Lexical Perspectives on Transitivity and Ergativity: Causative Constructions in English

LEXICAL PERSPECTIVES ON TRANSITIVITY AND ERGATIVITY AMSTERDAM STUDIES IN THE THEORY AND HISTORY OF LINGUISTIC SCIENCE...

Author: Maarten Lemmens

195 downloads 1059 Views 14MB Size Report

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

DOWNLOAD PDF

LEXICAL PERSPECTIVES ON TRANSITIVITY AND ERGATIVITY

AMSTERDAM STUDIES IN THE THEORY AND HISTORY OF LINGUISTIC SCIENCE General Editor E. F. KONRAD KOERNER (University of Ottawa) Series IV - CURRENT ISSUES IN LINGUISTIC THEORY

Advisory Editorial Board Henning Andersen (Los Angeles); Raimo Anttila (Los Angeles) Thomas V. Gamkrelidze (Tbilisi); John E. Joseph (Edinburgh) Hans-Heinrich Lieb (Berlin); Ernst Pulgram (Ann Arbor, Mich.) E. Wyn Roberts (Vancouver, B.C.); Danny Steinberg (Tokyo)

Volume 166

Maarten Lemmens Lexical Perspectives on Transitivity and Ergativity Causative constructions in English

LEXICAL PERSPECTIVES ON TRANSITIVITY AND ERGATIVITY CAUSATIVE CONSTRUCTIONS IN ENGLISH

MAARTEN LEMMENS Universite Charles de Gaulle, Lille 3

JOHN BENJAMINS PUBLISHING COMPANY AMSTERDAM/PHILADELPHIA

The paper used in this publication meets the minimum requirements of American National Standard for Information Sciences — Permanence of Paper for Printed Library Materials, ANSI Z39.48-1984.

Library of Congress Cataloging-in-Publication Data Lemmens, Maarten. Lexical perspectives on transitivity and ergativity : causative constructions in English / Maarten Lemmens. p. cm. -- (Amsterdam studies in the theory and history of linguistic science. Series IV, Current issues in linguistic theory, ISSN 0304-0763 ; v. 166) Revision of the author's thesis (doctoral). Includes bibliographical references (p. ) and index. 1. Grammar, Comparative and general-Verb. 2. Causative (Linguistics) 3. Semantics. 4. English language-Transitivity. 5. English language--Ergative constructions. I. Title. II. Series. P281.L46 1998 415--dc21 98-36270 ISBN 90 272 3671 2 (Eur.) / 1 55619 882 5 (US) (Hb; alk. paper) CIP © Copyright 1998 - John Benjamins B.V. No part of this book may be reproduced in any form, by print, photoprint, microfilm, or any other means, without written permission from the publisher. John Benjamins Publishing Co. · P.O.Box 75577 · 1070 AN Amsterdam · The Netherlands John Benjamins North America · P.O.Box 27519 · Philadelphia PA 19118-0519 · USA

To Hilde Andreas, Joanna, Simon, and Jan-Willem with love and apologies

ACKNOWLEDGMENTS The present book is an extensive revision of my doctoral dissertation. While many of the analyses and observations go back to the original work, the present book can safely be said to be a complete rewrite of the original manuscript. The theoretical descriptions have been pruned to what is absolutely essential and the descriptive material has been restructured to better clarify my own cognitive lexical-paradigmatic approach to causative constructions in English, basically a synthesis of cognitive grammar and systemic-functional grammar. It is my pleasant duty to thank the many people who, in one way or another, have contributed to both my doctoral dissertation and this book. It is impossible to list all these people, yet I would like to thank some of them explicitly. I wish to thank Dirk Geeraerts, promotor of my Ph.D. dissertation, for having continually challenged and stimulated me with his methodological concerns and his persistently careful criticism. While less directly involved in the rewriting process, his comments on first drafts of this book have been much appreciated. I am grateful to my Brygida Rudzka-Ostyn, co-promotor of my Ph.D. dissertation, for her elaborate comments on both the contents and the presentation of my work. Her extensive feed-back, energy and enthusiasm have been a major impetus to my work. I also thank Kristin Davidse for her careful comments on substantial parts of this work, which helped me considerably in the delineation of my own approach. I am indebted to the department of Linguistics of the K.U. Leuven for offering me an assistantship during which I carried out most of the linguistic research reported on in this book. I thank Emma Vorlat for granting me the academic freedom for this research. The colleagues at the department provided a pleasant and stimulating atmosphere. The time for research granted by the Université Charles de Gaulle, Lille, France, where I work at present has been greatly appreciated. I thank my colleagues in the English department for their interest and support. I am much obliged to Paul Bijnens, Jan Ducaju, Fred Truyen and Ivo Jossart for their help with the computer-part of this work. I thank Dirk Speelman for providing me with citations from the OED on CD-ROM. In the course of the years I have had the opportunity to meet several people and discuss my work with them. I particularly thank Eugene Casad, Seana

viii


Coulson, Adele Goldberg, Louis Goossens, Ismael F. Hussain Al-Bajjari, Andreas Jucker, Dan Jurafsky, Robert Kirsner, Philip Miller, Peter Norvig, Mary-Ellen Ryder, David Tuggy and Dekai Wu for their constructive comments or their stimulating interest in my work. I thank my proofreaders Adrian Gebbett, Camilla Mryglod, Gert Troch and Arthur Gamble. Responsibility for the final product is, of course, mine. I am much endebted to the two anonymous reviewers appointed by John Benjamins. Their constructive comments have helped me considerably in streamlining my work and rework it into the present book. I thank Anke de Looper from Benjamins for her help with the formatting of this book and other practical matters. A special word of thanks I reserve for my best friend Kurt Feyaerts. He shared with me the ups and downs of both my Ph.D. and this book. With a faith in me often larger than my own, he never stopped encouraging me to continue with my work. Great is my debt to my family. Warm thanks go to my parents and parentsin-law for their support and concern, and various kinds of practical help. To my wife Hilde I am most indebted for her love, her optimism and ability to put things in perspective. She moreover has the gift of performing miracles and making things work where nobody else could or would. Without her, this work would simply not have been possible. Andreas, Joanna, Simon, and Jan-Willem made an essential contribution by their heart-warming spontaneity, their overwhelming joy of life, and particularly by limiting their interest in this book to an inquiry into the number of pages and when it would be finished. I am glad I can tell them I am home again.

CONTENTS

Part I: Theoretical Premises Chapter 1: Introduction 1.1 Constructional variation with causative verbs 1.2 Structural organization of this book 1.2.1 Part I: Theoretical premises 1.2.2 Part II: Case studies 1.3 Theoretical framework 1.3.1 Cognitive Grammar 1.3.2 Systemic Functional Grammar 1.3.3 Generative and Relational Grammar 1.4 Empirical methodology 1.4.1 Advantages of corpus-based research 1.4.2 Corpora consulted

3 5 5 6 8 8 13 14 17 17 18

Chapter 2: The semantics of causative constructions 2.1 Analytical versus lexical causatives 2.2 Transitive versus ergative lexical causatives 2.2.1 Traditional form-based transitive/ergative typology 2.2.2 Langacker's cognitive approach 2.2.3 Levin's lexical-semantic approach 2.2.4 Davidse's paradigmatic approach 2.2.5 Present cognitive lexical-paradigmatic approach

22 28 28 30 36 39 46

Chapter 3: The semantics of causative verbs 3.1 The structure of lexical categories 3.2 The structure of lexical fields 3.3 Lexical versus constructional semantics 3.4 Conclusions to Part I

49 55 62 66

χ


Part II: Case Studies Chapter 4: Lexically determined flexibility and constraints 4.1 Introduction 4.2 A synchronic perspective on constructional constraints 4.2.1 Defining the middle construction 4.2.2 A lexical-paradigmatic view on the middle construction 4.2.2.1 Transitive versus ergative middle constructions 4.2.2.2 Lexical constraints on the middle alternation 4.2.2.3 English middable verbs 4.3 A diachronic perspective on constructional flexibility 4.3.1 The ergativization of intransitives 4.3.1.1 Starve 4.3.1.2 Hunger and Thirst 4.3.1.3 Ergativization as linguistic creativity 4.3.2 The ergativization of transitives 4.3.3 Complex patterns of paradigm shifts 4.3.3.1 Throttle an engine 4.3.3.2 Hang 4.3.3.3 Abort 4.4 Conclusions

71 71 72 74 74 78 82 85 86 87 89 91 92 94 94 96 96 97

Chapter 5: The Agent-centredness of the transitive paradigm 5.1 The experiential basis of agentivity 5.2 Lexical blueprints for agentivity: the MURDER verbs 5.2.1 Overview of the MURDER verbs 5.2.2 Murder 5.2.3 Assassinate 5.2.4 Execute 5.2.5 Lynch 5.2.6 Slay 5.2.7 Massacre 5.2.8 Butcher and Slaughter 5.2.9 Kill 5.2.10 Causes and Agents 5.3 Morphological blueprints for agentivity: agentive nominals 5.3.1 The semantic network for -er 5.3.2 A lexical-paradigmatic view on -er 5.3.2.1 Transitive preference of -er 5.3.2.2 Paradigmatic complexities of-er 5.3.3 Summary

99 106 106 107 108 109 111 111 113 113 115 123 127 128 131 131 135 139

CONTENTS

xi

5.4 Constructional blueprints for agentivity: the objectless transitive 5.4.1 A lexical view on the objectless transitive 5.4.2 A lexical-paradigmatic view on the objectless transitive 5.5 Conclusions

140 140 141 146

Chapter 6: The Medium-centredness of the ergative paradigm 6.1 The experiential basis of ergativity 6.2 Lexical and constructional convergence in Early Modern English 6.3 Lexical and constructional differentiation in Contemporary English 6.3.1 Overview of the SUFFOCATE verbs 6.3.2 Suffocate 6.3.3 Asphyxiate 6.3.4 Smother, Smoulder, and Smore 6.3.5 Drown 6.3.6 Stifle 6.3.7 Strangle and Throttle 6.3.8 Choke 6.3.9 Summary 6.4 Transitivity and force-dynamics: phrasal verbs 6.4.1 Overview of constructional deviations 6.4.2 Choke down 6.4.3 Choke back 6.4.4 Choke out 6.4.5 Choke off 6.4.6 Choke up 6.4.7 Drown out 6.4.8 Summary 6.5 Conclusions

149 153 158 158 158 159 160 161 164 165 168 173 178 178 178 179 180 181 183 184 186 187

Chapter 7: Complexities of the transitive-ergative interplay 7.1 The experiential basis of the transitive/ergative interplay 7.2 Lexical and constructional variation in Modern English 7.3 Ergativization of abort in Early Modern English 7.4 Transitivization of abort in Modern English 7.5 Conclusions

191 192 197 208 216

Chapter 8: Summary and evaluation 8.1 Summary 8.2 Evaluation 8.2.1 Lexical and constructional interdependency 8.2.2 The cognitive reality of transitivity and ergativity 8.3 Further Research

221 231 231 233 238

xii


References Dictionaries Corpus References Glossary Index of Names Index of Subjects

243 255 256 257 259 262

PART I

THEORETICAL PREMISES

CHAPTER 1 INTRODUCTION

1.1 Constructional variation with causative verbs This work ultimately argues that in order to explain the variable behaviour of causative verbs in modern English, it is fundamental to consider not only the lexical properties of those verbs but also the interaction between those properties and the meanings of the constructions in which the verbs may appear. My corpus-based analysis of verbs of killing, verbs which in linguistic treatises are generally taken as causative verbs par excellence, reveals the complex conceptual networks of (related) meanings instantiated by these verbs. With respect to constructional meaning, I continue, and by doing so refine, the view defended in Davidse (1991, 1992) that the English grammar of causative constructions is governed by the transitive and ergative paradigms. Overall, my perspective on the relevant grammatical constructions remains a cognitively inspired lexical one, aiming at clarifying how lexical and constructional meaning interact dynamically. The linguistic phenomena to be addressed in this book can best be identified via some seemingly puzzling questions related to expressions of killing: - Why can one rephrase John choked Mary (a causative) as Mary choked (a non-causative), while one cannot rephrase John killed Mary as *Mary killed ? - The same type of rephrasing is acceptable for drown in the sentence The former government drowned hundreds of dissidents (vs. Hundreds of dissidents drowned after being dumped into the ocean from a plane). However, it is not acceptable in The protesters ' horns drowned out the bell (*The bell drowned out). Why ? - Why does murder allow an objectless construction with the agent in subject position, as in, They have the clear intention to murder, whereas in a comparable objectless construction with starve, e.g. They have the clear intention to starve, the subject cannot be interpreted as the agent causing someone else's starvation but as the entity that is affected by it ?

4


- Why is it that many causatives can occur in a medio-passive construction, e.g. This book reads easily or These glasses break easily whereas such constructions are at least awkward for verbs of killing, e.g. *?Mary kills easily (with Mary being the victim). - Why is it that abort allows the causative/non-causative alternation in metaphorical uses (e.g. the pilot aborted the takeoff vs. the takeoff aborted), but does not allow it for literal uses (e.g. the woman aborted the child vs. *the child aborted) ? While these questions are mere appetizers, they all relate to a verb's constructional variability coupled with its distinct meanings. To adequately answer these and related questions concerning the interaction of verb meaning and constructional variability, it is necessary to characterize verbs of killing against the background of the transitive and ergative paradigms. In contrast to, for instance, some Asian or Amerindian languages, English does not formally mark the ergativity/transitivity distinction (cf. Dixon 1979, 1994; Comrie 1981 or DeLancey 1981, 1984a, 1990 for illustrations of such markings). As a result, many scholars have failed to observe it as a principle operative within the grammar of English. However, can one claim at all that the English grammar of causation is structured by the ergative and transitive systems if the distinction between the two is not marked overtly, e.g. by case markers? Following Davidse (1991, 1992), I argue that the claim stands up to close examination, provided that one recognizes that overt markings are but one way in which the two systems manifest themselves in a language. I accept the Whorfian view that overt grammatical categories should be distinguished from covert ones which make themselves felt by the systematic relations that link all the paradigmatic correlates of a construction type. The semantics of a covert grammar can be arrived at by systematically examining the paradigmatically related constructions. This type of heuristics has been exploited insightfully within the framework of systemic-functional grammar as, for instance, in Davidse's (1991) innovative work on transitivity and ergativity in the English grammar. Within mainstream linguistics Levin's (1993) work on English verb classes is noteworthy. Levin effectively uses the alternation method to characterize a verb's meaning. The major shortcoming of her work is that she sees the choice of constructions in which a verb may occur as wholly determined by the verb's semantics and, as a result, fails to recognize that the constructions themselves are meaningful (see Chapter 2, section 2.2.3 for more explicit criticism). That constructions have meaning independent of verbs is argued, for example, in Goldberg (1995). While she discusses construction types different from

INTRODUCTION

5

the ones analysed in this work, her views are largely compatible with mine. Constructions are meaningful in and of themselves, but as Goldberg rightly insists, "it is clearly not the case that the grammar works entirely top-down, with constructions simply imposing their meaning on unsuspecting verbs" (1995:24). Rather, she says that "the meaning of constructions and verbs interact in non-trivial ways" (1995:24). The interaction between verbal and constructional semantics is indeed a complex matter. Within the scope of the present work, it can be restated in terms of the following questions: - How does the lexical content of a verb influence constructional variability? - Conversely, how do the various constructions in which a verb may occur affect its meaning? - What is the role of extension mechanisms like metaphor and metonymy with respect to the ergative-transitive interplay? - How may lexical and constructional properties change over time and does this lead to prototype shifts? My aim is to suggest some answers to these complex and far-reaching questions by analyzing a large corpus of verbs of killing. 1.2 Structural organization of this book 1.2.1 Part I: Theoretical Premises The first part of this book consists of three chapters which present a more general and theoretical exploration into some issues that are immediately relevant to my overall argument. As the first of these three chapters, the present chapter has stated the purpose of this book as essentially involving an account of constructional variability with lexical causatives. It further clarifies the general theoretical assumptions underlying this work (section 1.3), which are in essence an innovative synthesis of (predominantly) Cognitive Grammar and Systemic-Functional Grammar. The chapter concludes with a clarification of the strong empirical basis of our analysis (section 1.4). The next two chapters of Part I deal more elaborately with the grammatical and lexical perspectives which form the backbone of my work. Chapter 2 presents a careful description of causative constructions. First, I argue how lexical and analytical causatives, which have often been regarded as semantically equivalent, in fact involve quite different conceptualizations of an event (section 2.1). The second and most important part of Chapter 2, section 2.2, is concerned with a proper characterization of lexical causatives. It offers brief critical discussions of some views which have become accepted within

6


certain schools of thought and which in one way or another are relevant to the theoretical synthesis underlying the present work. This synthesis, a cognitive lexical-paradigmatic approach, is characterized briefly in the last section of Chapter 2 (section 2.2.5); its descriptive and explanatory adequacy is fleshed out to the fullest in the descriptive analyses in Part II of this book. Chapter 3 takes up the other important pillar of my work, viz. the lexical semantic perspective on (causative) verbs. Clearly, the lexical perspective focuses on those aspects of verbal meaning that directly pertain to the verb's constructional potential. The chapter clarifies the cognitive view on semantic multiplicity, which I hold as the most appropriate to account for semantic variability both in lexis and in grammar. Following Langacker (see References), it is argued that conceptual categories are not rigidly and absolutely defined structures, but prototypically structured networks of interrelated meanings. It will be shown that this not only holds for lexical categories as for example instantiated by individual verbs of killing (section 3.1), but also for lexical fields, such as that designated by the group of verbs of killing (section 3.2). The next section in the chapter offers a preliminary clarification of the relationship between constructional variability and lexical meaning relatedness. The concluding section in Chapter 3 is in fact the conclusion to the whole of Part I (note that neither Chapter 1 nor Chapter 2 have a conclusion) and is at the same time an opener into the descriptive analyses in Part II. 1.2.2 Part II: Case Studies Part II of this book, comprising the actual descriptive work, contains four chapters which have been set up to cogently describe the complexities of the transitive and ergative models and their dynamic interaction with verbal meaning. Chapter 4 focuses on how the lexical structure of verbs may either constrain or encourage constructional extensions. Following a brief introduction, the first analysis (section 4.2) deals with lexically determined constraints on the middle or medio-passive construction in present-day English. Ironically perhaps, the insights on the middle construction originated in the need to account for the unmistakable absence of this construction for verbs of killing. The description augments Langacker's cognitive analysis by a paradigmatic perspective, yet at the same time remedies some shortcomings of Davidse's paradigmatic views on this construction. The second description offered in Chapter 4 (section 4.3) is a diachronic perspective on different types of paradigmatic shifts within the field of killing. It shows how over a considerable time span, the lexical evolution of some verbs, e.g. starve or throttle, has taken the verbs into new constructional possibilities.

INTRODUCTION

7

Chapter 5 deals with the Agent-centredness of the transitive paradigm, and explores this characteristic from different perspectives. First, it considers the prototype structure of the Agent-category and shows how agentivity is experientially motivated (section 5.1). Next, reviewing the individual MURDER verbs, it reveals how the Agent-centredness of this most salient (and prototypically transitive) subset of the field of killing is lexically determined (section 5.2). In a lexical-paradigmatic analysis of agentive nominals in -er (e.g. killer), section 5.3 explores the Agent-centredness of the transitive paradigm from a morphological perspective. Finally, our analysis of the transitive construction with omitted object further elucidates, from a constructional point of view, the fundamental hook-up with agentivity that characterizes the transitive paradigm. As a logical complement to Agent-centredness of the transitive paradigm, Chapter 6 shows how the ergative paradigm is primarily concerned with the participant affected by a process. By discussing a group of predominantly ergative verbs, the SUFFOCATE verbs, the chapter further reveals the experiential basis of ergativity for these verbs, yet it also shows how the opposition between transitive and ergative SUFFOCATE verbs is experientially motivated as well. The rest of the chapter reviews the enormously rich semantic and constructional coverage of the SUFFOCATE verbs. Section 6.2 elaborates the brief diachronic description of Chapter 4, section 4.3.2, and further substantiates the lexical and constructional overlapping that characterizes the history of these verbs. Next, section 6.3 (which forms the logical counterpart to section 5.2 in Chapter 5) describes the reorientation towards more differentiated lexical and constructional prototypes underlying the evolution to the present-day situation. The final section of the chapter discusses specific cases of transitivization' of ergative choke and drown triggered by particles. Chapter 7 provides a maximal focus on the complexities of the interplay between the transitive and ergative paradigms, elaborating on the paradigmatically mixed character of the verb abort. By tracing the complex etymological evolution of this verb, we can account for the verb's constructional complexity in present-day English. It finds a logical explanation in view of the symbiosis of lexical and constructional meaning. By looking at grammatical structures from the lexical end, the different chapters in Part II clarify how transitivity and ergativity operate within the field of killing and as such may provide a useful basis for extrapolations to other types of causative verbs in the English lexicon. Starting from a general overview of my findings, the final conclusions to this work further reflect on the nature of verbal and constructional interaction and sketches some avenues for further research.

8


To facilitate the comprehension of the text, an index and a glossary of terms are added at the end of this book. The glossary briefly defines the terminology used in this work and presents a list of corpus references used (see also section 1.4 below). 1.3 Theoretical framework As to its underlying assumptions, the present work adheres primarily to the principles of Cognitive Grammar as developed by Langacker (see References), especially in its treatment of lexical meaning. In addition, this study draws considerably on Systemic-Functional Grammar as developed by Halliday and others (especially Davidse 1991) whose views are in many respects strikingly compatible with those of Cognitive Grammar. At the same time, this work incorporates some descriptive insights from work in the framework of Generative Grammar and Relational Grammar and their derivatives (e.g. Keyser & Roeper 1984, Levin 1993; Levin & Rappaport Hovav 1995). As such, the theoretical underpinnings of this work offer what I believe to be an innovative synthesis of these frameworks. While the result may no longer be fully compatible with rigid formulations of any of these, it offers insights which can be pertinent to all. The following sections briefly map out the major assumptions of these frameworks. The discussion, necessarily very brief, starts with Cognitive Grammar, the theoretical homeground of this book. Next, it discusses some notions of Systemic-Functional Grammar from which the insights concerning the paradigmatic opposition between transitivity and ergativity are borrowed. Finally, some aspects of Generative and Relational Grammar deserve some attention. It should be stressed that these descriptions serve to situate the present work in the linguistic tradition and do not aspire to be a full-fledged evaluation of these models. More critical evaluations occur at various places in this book, when specific views (e.g. on analytical versus lexical causatives or on agentivity) are at issue. 1.3.1 Cognitive Grammar Within the framework of Cognitive Grammar, there is no principled distinction between lexicon and grammar, both of which form a continuum of symbolic structures. More precisely, grammar is defined as a structured inventory of conventional linguistic units, i.e. structures which "the speaker has mastered thoroughly to the extent that he can employ [them] in largely automatic fashion, without having to focus his attention specifically on [their] individual parts or their arrangement" (Langacker 1987a:57). Three types of units are distinguished: semantic, phonological, and symbolic. The latter are bipolar and as such associate a phonological structure with a semantic structure.

INTRODUCTION

9

To say that lexical and grammatical structures form a continuum does not mean that they are identical; grammatical structures are more schematic than lexical ones. A sentence like The dog / bit / the postman combines three meaningful units into a larger structure which itself is also meaningful as it represents the sequence AGENT-PROCESS-AFFECTED.1 What this grammatical structure means is that some Agent acts upon another entity, the Affected. In this particular example, the identity of the Agent and the Affected are specified by the lexical items dog and postman, and the action by bite. The definite articles and the tense and voice of the verb are also regarded as meaningful. Langacker's symbolic units are equivalent to the constructions in Construction Grammar (see e.g. Fillmore 1986, 1988; Fillmore et al. 1988; Kay 1990 or Goldberg 1992, 1995). Like Cognitive Grammar, Construction grammar does not assume a strict division of lexicon and syntax. As Goldberg puts it, "both lexical and grammatical constructions are essentially the same type of declarative data structure: both pair form with meaning" (1995:7). I will be using the term 'construction' more restrictively, viz. in reference to the grammatical frames in which a verb may occur. For instance, the government starved its own children realizes an ergative effective construction (INSTIGATOR-PROCESS-MEDIUM), whereas the Ethiopians starved realizes the ergative non-effective construction (MEDIUM-PROCESS). These two constructions are formally and semantically different: the effective one overtly codes the entity that initiated the given event, while the non-effective remains neutral as to whether it was externally or internally instigated (see Chapter 2, sections 2.2.4 and 2.2.5, for more details). Returning to Cognitive Grammar, symbolic units are seen as providing the means for expressing conceptualizations in linguistic form. Cognitive grammar thus equates meaning with conceptualization. That is, semantic structure is defined as conceptualization "tailored to the specifications of linguistic convention" (Langacker 1987a:99). The meaning of a linguistic expression is a cognitive structure characterized relative to cognitive domains, "where a domain can be any sort of conceptualization: a perceptual experience, a concept, a conceptual complex, an elaborate knowledge system, etc." (Langacker 1991a:3). As Lakoff (1987) has shown, most of these are idealized cognitive models (ICMs). Such models are similar to Fillmore's frames, defined as "unified frameworks of knowledge, or coherent schematizations of experience" (Fillmore 1985:223).

1

In fact, the nominals the dog and the postman are themselves already complex units, but I decide to ignore this here.

10


In this view, linguistic semantics is encyclopedic in nature, involving specifications in many cognitive domains.2 The traditional distinction (cf. Lyons 1977) between linguistic knowledge and encyclopedic knowledge is thus untenable. As Haiman (1980:331) states it, "dictionaries are encyclopedias" (cf. also Langacker 1987a:154ff, Croft 1993). The relevant background information for the characterization of meaning constitutes "a network of shared, conventionalized, to some extent perhaps idealized knowledge, embedded in a pattern of cultural beliefs and practices" (Taylor 1989:23).3 Against the background of such larger conceptual structures, called the 'base', linguistic structures impose their own specifications, which brings us to one of the pivotal claims of Cognitive Grammar, viz. that "'linguistic expressions and grammatical constructions embody conventional imagery" (Langacker 1988a:7). Meaning thus relies on our ability to conceptualize the same object or situation in different ways. As Casad summarizes it: "the speaker's ability to conceptualize situations in a variety of ways is, in fact, the foundations of cognitive semantics" (1995:23). The following chapters will provide ample illustration of different dimensions of imagery; by way of illustration, I elaborate here only one, viz. 'figure/ground' alignment. This dimension involves a basic principle of cognitive and perceptual experience by virtue of which humans perceive (or cognize) entities as standing out against others. The meaning of linguistic units can also be seen in terms of figure/ground alignment, viz. as the imposition of a profile on some relevant base. The profile is "a substructure that is elevated to a special level of prominence within the base" (Langacker 1991a:5). Different predications can be characterized relative to one and the same domain, yet as they normally diverge in their figure/ground alignment, they are semantically non-equivalent. As an example, consider the relationship between the verb kill and its nominalization killing in the following examples: (1)

Yusef Hawkins was killed last week ... (WSJ)4

2 See Lakoff (1987:75ff) for an illustration of the complexity of conceptual matrices (or in his terms, Idealized Cognitive Models or ICMs). 3 Gruber (1985) insists on the distinction between linguistic and encyclopedic meaning, and in fact complicates matters by adding yet another, viz. 'conceptual meaning'. In his opinion, "lexical and conceptual meaning are essential parts of the meaning of a word or the criteria for its use [...] Encyclopedic meaning is knowledge or beliefs merely associated with a word or category, i.e. inessential for its use" (1985:255; my emph.). Gruber introduces these distinctions to explain degrees of semantic acceptability (or anomaly). These gradiences find a more natural account within an encyclopedic view on meaning as advocated here (see the cited works by Croft, Haiman, Lakoff, Langacker and Taylor.) 4 The reference serves to identify the source of the example. See section 1.4 below for details.

INTRODUCTION

(2)

11

... the killing of a black youth by whites last month in Brooklyn. (WSJ)

Both examples refer to the same event, and from a truth-conditional perspective they would be considered semantically equivalent. The difference would be merely one of form. Within the framework of cognitive grammar, however, the verbal and nominal forms do not have the same meaning since they impose different profiles. They have the same conceptual base but differ with respect to how this base is treated. (I ignore some additional differences (e.g. the semantic import of the by and of-periphrasis) since they are not essential to the present discussion.) To appreciate the differences, it may be useful to take a closer look at profiles imposed by verbs and nouns. Verbs are relational predications that profile a temporal sequence of interconnections between entities. Langacker identifies such a state-by-state profile as sequential scanning. This 'close-up' view stands in polar opposition to the holistic view an episodic noun imposes on the same base. Such a noun is characterized by summary scanning. It profiles the component states collectively as a thing, i.e. as "a region in some domain, where a region is characterized abstractly as a set of interconnected entities" (Langacker 1987a:214).5 In the case of an episodic noun, the process functions as the base for the predication: the temporal and relational dimensions are still present but are not in profile. Figure 1, based on Langacker (1987a:247), contrasts the profiles that define the two grammatical categories (where profile is indicated via the thickness of the lines). event (verb) (a)

reification (nominalization) (b)

Fig. 1 : Process (a) and reified process (h)

5

As is obvious, the term 'thing' is by no means limited to physical objects, although the latter are probably typical instantiations of the category. The term is used as an abstract technical term to refer to the product of any conceptual reification. See Langacker (1987a) on the difference between sequential and summary scanning.

12


Figure 1(a) represents a processual expression which profiles the relationship between the participants (represented by the two circles) in its evolution through conceived time (indicated by the heavy-line segment of the time arrow). Figure 1(b) represents a nominalization as a "conceptual reification of an event conception" (Langacker 1992:497). It profiles the abstract region of interconnected states. To keep things manageable, the simplified representation shows only three states of what in fact is an indefinite succession of states. Relational predications, central to this book, are more complex than nouns since they have a complex profile that also manifests figure/ground (= 'trajector/landmark') alignment. Within the relationship represented by Figure 1(a), the trajector (tr) is identified as the primary participant, the landmark (lm), as the secondary participant. The passive construction is thus characterizable as a figure/ground reversal — "the switching of what is foregrounded in a given scene with what is backgrounded" (Emanatian 1993; cf. also Langacker 1987a:120ff). To return to the basic assumptions of Cognitive Grammar, the grammar of a language is, as said, a "structured inventory of conventionalized linguistic units" (Langacker 1991b:511). All grammatical structures, be they morphological, lexical or syntactic, emerge as inherently meaningful. For example, while many grammars regard particles as meaningless, the description of choke {back/down/off/up/out} and drown out in Chapter 6, section 6.4, demonstrates their semantic contribution to the clause meaning and their impact on the verb's constructional possibilities. The constructions themselves are inherently meaningful too. Constructionally related pairs like John killed Mary versus John killed or Burglars murdered Jill versus Jill was murdered by burglars are semantically non-equivalent because they serve to conceptualize the event differently. In plain terms, the cognitive dictum is that any difference in form entails a difference in meaning (cf. e.g. Wierzbicka 1988 or Langacker 1985, 1990b). Naturally, a given reality can be conceptualized in different ways. Chapter 6 argues that the lexical and constructional overlapping in the group of SUFFOCATE verbs is in fact largely due to subtle variations in the conceptualization of what is ontologically quite similar. An important characteristic of Cognitive Grammar is the view on (semantic) categories as not always well-delineated and structured around a prototype, the conceptual centre of the category. A category may have peripheral members, which are defective in one way or another; often these peripheral members lie in the blurred transition area between two categories. Langacker analyzes the members of a category as "nodes in a networks, linked to one another by various sorts of categorizing relationships" (1991a:266). Of these relationships, 'instantiation' and 'schematization' are particularly relevant, as it allows one to unite under one more abstract and general structure, called a 'schema', various

INTRODUCTION

13

structurally more varied lower-level constructions, the instantiations. A more elaborate discussion of Langacker's model of categories is offered in Chapter 3, section 3.1. 1.3.2 Systemic Functional Grammar In many ways cognitive grammar is compatible with Halliday's systemicfunctional grammar. "A language", Halliday says, "is a system for making meanings" (1985:xvii). As in Cognitive Grammar, the term 'semantic' is not reserved for the lexicon, but pertains "to the entire system of meanings of a language, expressed by grammar as well as by vocabulary" (1985:xvii). As also observed by Wierzbicka, in language, "everything 'conspires' to convey meaning" (1988:2). Parallel to the cognitive grammar views on the conceptualization of reality, Halliday emphasizes that the semantic structures of a language "enable us to 'think about' our experience—that is to interpret it constructively" (1985:xviii). Semantic structures being lexical and syntactical, Halliday thus commits himself to the view that syntax is motivated by meaning. However, an important difference with cognitive grammar (and other frameworks) is that Halliday sees the motivation as a 'natural' (i.e. symbolic) relation between meaning and lexicogrammar (lexis and syntax) and that is why language is learnable: "[children] can make the link between the categories of the grammar and the reality that is around them and inside their heads" (1985:xviii). This 'plausibility' is for instance reflected in the fact that English has verbs and nouns, "to match the analysis of experience into processes and participants" (1985:xviii). This congruence with experience can, however, be overturned by a process of grammatical metaphor, via which categories can be cross-coded (e.g. a process represented by a nominal phrase). Against the background of the previous discussion of Cognitive Grammar, the compatibility between Halliday's and Langacker's view will be appreciated. Instead of 'congruence', Cognitive Grammar uses the notion 'canonical conception' which, too, is seen as having a strong experiential basis. While the difference between Halliday's tri-stratal view of the linguistic sign (meaning - lexicogrammar - phonology) and the bipolar symbolic units of Cognitive Grammar calls for more careful analysis, it will not be elaborated in the present work, as it does not hamper the successful integration of the descriptive merits of the two models. Following Hjelmslev, Halliday characterizes language as "system and process" (1985:xxii). The system is to be seen as a highly complex network of interrelated choices: "it is a paradigmatic, non-linear meaning potential" (Davidse 1987:46). With the notion "language as process", Halliday refers to actual instances of language use (which he calls 'text'), which is "linear and

14


syntagmatic" (1987:46). The paradigmatic nature of systemic grammar means that "describing something consists in relating it to everything else" (Halliday 1985:xxvii; his emph.). It is precisely by tracing the paradigmatic organization of language (e.g. via the alternation method) that relevant insights can be obtained regarding the grammar of actions and events. Up to the present, linguistic analyses in Cognitive Grammar have continued to focus primarily on the syntagmatic side and neglected the paradigmatic; this work expands Cognitive Grammar to include the latter into its analyses. Clearly, the paradigmatic options are constrained, "not only from the grammar itself, but also from the context and the culture" (Davidse 1987:46). This view is quite compatible with the cognitive view on linguistic construal. Next to language as 'system', equal weight is given within SystemicFunctional Grammar to language as process, where process is seen as 'text', i.e. any actual instance of language use in context. As Halliday points out, one must carefully consider how the system engenders text and at the same time, text should be related to the system behind it, "since anyone understanding the text does so only because they know the system" (1985:xxii). The present work will also have a strong empirical basis (cf. section 1.4 below), which allows us to come to nuanced judgments on how the paradigmatic contrast between transitivity and ergativity operates within English. A strong concern with data (language in context), as opposed to idealized language systems, is also a feature of Cognitive Grammar, characterized by Langacker as a usage-based model. In such a model, next to a concern for general rules and principles, "we must also give substantial weight to their arrays of conventional instantiations, investigating the actual extension of the patterns in question and the factors that influence it" (Langacker 1991a:265). All in all, despite some fundamental differences between systemic and cognitive linguistics—notably, Halliday's multi-layered approach to language structure and his emphasis on the 'interpersonal' dimension of grammar—and differences of perspective or terminology, there is considerable affinity between the Systemic-Functional Grammar and Cognitive Grammar. Exploring the metalinguistic divergencies and similarities between Systemic-Functional Grammar and Cognitive Grammar will surely be a profitable enterprise; however, as these are not directly relevant to the present work, they will be disregarded. The present study tries to break through the metalinguistic shells of different frameworks and distil from them their descriptive value. 1.3.3 Generative and Relational Grammar The theoretical assumptions of Generative and Relational Grammar (and their derivatives) are radically different from those of Cognitive Grammar and Systemic-Functional Grammar. A major criticism against Generative and Rela-

INTRODUCTION

15

tional Grammar is that they assume syntax to be autonomous from semantics. Hence, they often claim relationships between structures (e.g. active and passive or nominal and verbal constructions), but see these as semantically equivalent. In both Cognitive Grammar and Systemic-Functional Grammar, in contrast, semantics forms the basis for grammar. Langacker stresses that "grammatical structure reduces to patterns for the structuring and symbolization of conceptual content, and that all valid grammatical constructs have some kind of conceptual import" (1991a:338-9). Similarly, Halliday points out that paradigmatic grammars "take semantics as the foundation" (1985:xxviii). In this view, any structure is meaningful and consequently, the notion of meaningless structures (dummies or 'chômeurs') is rejected. As has already been emphasized above, the basic view is that any difference in form entails a difference in meaning. Logically, then, the division between syntax, morphology and lexicon is not strict, as all consist of meaningful units, admittedly of different 'levels of schematization' (Cognitive Grammar) or 'scales of delicacy' (SystemicFunctional Grammar). A systemic inspired point of criticism is that the Generative and Relational models are exclusively syntagmatic in orientation; they regard language as a list of structures between which relations may exist (hence the transformations or strata). Mostly, however, they fail to see the deeper system at work in the grammar and thus often make incorrect distinctions (if at all). With respect to the issues central to this book, they mostly miss the awareness of the transitivity/ergativity distinction. More specifically, they generally do not observe the distinction between a transitive two-participant construction (e.g. John killed Mary) and an ergative one (e.g. John suffocated Mary) nor do they distinguish between transitive and ergative one-participant constructions, e.g. Mary died vs. Mary suffocated (see Chapter 2, sections 2.2.4 and 2.2.5 on these distinctions). Another problematic issue is the view on language in terms of the distinction between competence and performance, which especially in the Chomskyan model figures prominently. Instead of positing a model which presupposes a hypothetical ideal speaker and which imposes the abstract structures of this ideal language onto language itself, Cognitive Grammar and SystemicFunctional Grammar start by observing language use itself, which they see as the analytic way to capture generalizations concerning the underlying semantics (cf. above).6 From the earliest formulations of his theory, Halliday has

6 The generative quest for general, preferably even universal, principles of grammar has tended to discourage the analysis of 'real life' data. As Gross (1976:861) observes, within the generative tradition, "motivated by a desire to treat linguistics at an abstract level, work based

16


incorporated the social dimension into his model, stressing that "language is not a self-sufficient entity, but that it has an essentially instrumental character: it is used in—and indeed evolved to serve—human interaction" (Davidse 1987:40). Cognitive Grammar, too, defends such a position. Despite these (fundamental) theoretical conflicts, much valuable research has been carried out within the Generative and Relational Grammar models. For example, the Unaccusative Hypothesis, as first introduced by Perlmutter (1978) and later taken over (and silently extended) by Government and Binding theory (e.g. Burzio 1986, Marantz 1984) has been concerned especially with different types of intransitives and drawn attention to certain differences, yet it is still flawed in many respects. Other work that has contributed to the insights of the present book are Horn (1980), Keyser & Roeper (1984), Levin (1993), Levin & Rappaport (1995). Levin's (1993) work is particularly interesting in two respects: (1) she explicitly couples a verb's constructional behaviour with its meaning and (2) she uses the heuristic of alternation to arrive at a proper characterization of the verb's constructional potential. As will be shown (Chapter 2, section 2.2.3), there a still some problems with her approach, one of them being her too rigid view on categories. Levin & Rappaport (1995) offer a more nuanced view which is considerably more compatible with those defended in this book. A brief evaluation is presented in the final conclusions of this book. In short, the present work wants to offer a synthesis of especially Cognitive Grammar and Systemic Functional Grammar and I believe that both models can greatly benefit from this synthesis. Although the tide is turning, Cognitive Grammar has had, since its earliest formulations in the late 70s, some bias towards lexical semantics (think only of the numerous preposition and metaphor studies, e.g. Lakoff & Johnson 1980, Brugman 1983, Hawkins 1984). This work hopes not only to emphasize that constructions themselves are meaningful units (a concern also central to e.g. Goldberg 1995), but also to offer new insights concerning the transitivity/ergativity distinction as operative in English grammar. Systemic-Functional Grammar, in turn, can profit from the lexical orientation of this work. In my view, too little attention has been paid within Systemic-Functional Grammar to the influence of the verbs' lexical structure on paradigmatic choices. It is in my emphasis on the lexically determined structural variability that my cognitive homeground manifests itself most clearly.

on sentences [...] has almost entirely vanished". In essence, what the generativists were after was an abstract linguistic theory, not a model of language usage.

INTRODUCTION

17

1.4 Empirical methodology 1.4.1 Advantages of corpus-based research Next to the theoretical framework, the methodology for data gathering and analysis greatly determines the conclusions of any descriptive study. While I do not want to overemphasize the practical decisions underlying this work, it is warranted that some of the choices be clarified. A major feature of the present work is its empirical foundation. I am committed to the idea that in order to gain insight into the principles of grammar, a close examination of a vast amount of contextualized, preferably non-elicited, language material is essential, for several reasons. First, a corpus-based analysis avoids inaccuracies and overgeneralizations which one often finds in intuitionbased studies. Secondly, it avoids the overconcentration on 'typical examples' which unambiguously illustrate whatever one is trying to prove. Extensive corpus material can shed light on less prototypical cases, leading to a better characterization of linguistic reality. Thirdly, corpus analysis may bring to the fore subtle semantic differences that would otherwise have been overlooked, as the analysis of collocational patterns for the SUFFOCATE verbs will prove.7 Clearly, a corpus-based analysis faces some shortcomings as well. First, it cannot with absolute certainty confirm the unacceptability of some or other construction, exceptions being explicit metalinguistic comments such as "John died Mary" is an ungrammatical English sentence.8 Nevertheless, provided that one's database is sufficiently large, the non-occurrence of a pattern may be held indicative of its status within the grammatical system. In addition, absence of certain constructions may trigger new insights into the semantics of the construction itself, as has been the case for our analysis of the middle construction which does not occur in the text corpora. A second drawback of corpus analysis is that there are practical limitations to the amount of material which can be analyzed within a certain time-frame. Although larger computerized corpora as well as more powerful tools for analysis are becoming more and more available, the work load is still considerable. As most corpus-based studies, the present work has also been delimited by what was practically feasible. Undoubtedly, such practical limitations prevent a study from reaching the ideal of 'maximal representativity'. While I find the notion of representativity itself quite challenging (cf. also Stubbs 1995), I have taken care to select data of different types in order to reach an acceptable degree of representativity. By doing so, certain

7

See e.g. also Sinclair (1991) or Stubbs (1995) for such collocation-based analyses. See Chapter 7 for an example of such meta-linguistic comment. Note that the problem of attestation also applies to diachronic research: the absence of a certain structure from historical documents does not necessarily mean that it has never existed. 8

18


reliable conclusions can be drawn from the data, indicating major tendencies and evolutions. 1.4.2 Corpora consulted More specifically, the corpora consulted comprise text from different registers, different regions (chiefly British versus American English), and different periods. Table 1 gives a short overview of the major (electronic) corpora used. Reference

Short Description

Size

WSJ

Wall Street Journal: computerized version of ±3 months of the paper

5,353,500 words

LDC

Leuven Drama Corpus: collection of 62 British plays from the late 60's and early 70's

1,029,660 words

CASS

Contemporary American Short Stories, taken from journals (Athene, Intertext, Quanta) distributed via the internet

1,066,875 words

NOV

Collection of some fifteen 19th century novels, distributed via the internet

746,525 words

Oxford English Dictionary (both paper edition and

11,713 citations

OED

CD-ROM)

Table 1 : Overview of corpora consulted Each of the corpora has its own specific character. The WSJ corpus, for example, stands somewhat apart because of its highbrow register and its 'journalese'.9 Nevertheless, it has been a rich resource for the MURDER verbs, given the many reports on various types of killings, murders, assassinations, and the like. The fiction corpora, in turn, have proven most rewarding for the SUFFOCATE verbs. Moreover, in comparison to the WSJ they show greater constructional variation; the reason is probably that in the fiction corpora, especially the CASS corpus, the creativity has not been hampered so much by newspaper editors (cf. Bell 1991) or by the formal enterprise of regular literary production under the influence of style-editors and market-conscious publishing houses.

9

I kindly acknowledge the Association of Computational Linguistics for granting me permission to use the ACL WSJ corpus.

INTRODUCTION

19

The Novel Corpus and the Oxford English Dictionary have been used to give some of our analyses diachronic depth. As for the OED it should be noted that I have consulted both the paper edition and the edition on CD-ROM. The latter made it possible to obtain any citation from anywhere in the dictionary. A specific reference to the CD-ROM version only occurs when the citation is not listed under the item's lemma in the paper edition. Whenever relevant, OED citations carry the date of attestation. When the citations deviates strongly from contemporary English, a gloss will be added. It is stressed that this gloss tries to reflect as much as possible the original phrasing, and no attempt is made to render a polished translation. The strong data-oriented character of the present work is reflected in the consistent use of (mostly corpus-based) examples to support my claims. It has also been my policy to vary the examples as much as possible. At all times I refer to the source of the example, using the abbreviations specified above. Examples drawn from other linguistic treatises will have a reference to these works. At times, I bring into the discussion some 'occasional' examples, i.e. ones that I have come across in my personal reading, that I heard on radio or television, or that I have taken from dictionaries. The reader is referred to the glossary at the end of the book for an overview of these occasional references. Examples without reference are my own or are self-made variations on attested usages. It should be noted that to enhance the readability of the text the examples have been slightly edited. Mostly, this involved leaving out irrelevant information. As a rule, the spelling has not been altered (e.g. the difference between American versus British spelling has been preserved). The database on which the present study is based has been arrived at by excerpting (via simple pattern-matching procedures) from the corpora all occurrences of kill-expressions. Chapter 3, section 3.2 gives a more elaborate and semantically based overview of expressions that are regarded as belonging to the field of killing. The excerpted attestations were subsequently marked with special codes to allow further analysis. It should be noted that the present work does not discuss all the extracted items in full detail. For instance, verbs of killing derived from instruments, such as to knife, to (machine-)gun or to poison are too few in number to yield significant conclusions. Expressions from semantically related fields, like die, lose one's life, perish, hunger or thirst are brought in for contrastive purposes. An overview of the items more elaborately discussed in Part II is presented at appropriate places in these descriptions. It should be noted that the extraction (and subsequent analysis) has not been restricted to verbs, but includes derived forms (e.g. killing, the kill, stranglehold, murderer, etc.) as well as related fonns which are not directly derived from the verb (e.g. abortion). In total, the number of attestations excerpted from the text corpora amounts to approximately 4000. Augmented

20


with those from the OED, the total of attestations examined adds up to over 15,000. The empirical basis of the present work has allowed us to come to more refined pronouncements on verbal and constructional meaning. As said, such a bottom-up approach furnishes insights into the complexities of lexicon and grammar from which general patterns and tendencies may be discerned. Clearly, while these generalizations are important, they are not the sole purpose of our investigation nor can they be regarded as telling the whole story in and by themselves. That is, they do not function as general rules from which all possible and well-formed linguistic structures can be predicted. Following Langacker (1987a, 1990b, 1991a), we reject such a one-sided quest for absolute predictability, which has figured prominently in Generative circles. Instead, we assume that abstract rule schemata coexist, and interact with, conventionalized instantiations. As Croft (1994:29) puts it, "[general] explanations are welcomed where found, but language is a mixture of functional motivation and arbitrary convention". Our investigation does support some predictability, e.g. the strong Agent-centredness of the MURDER verbs predicts that an ergativization will be barred. Nevertheless, even for seemingly clear-cut cases, absolute predictability is to be rejected. For instance, the Agent-centred SUFFOCATE verbs strangle and throttle have, at one point, been subject to ergativization. Our analysis shows that unexpected phenomena like these can be motivated on cognitive grounds. They can be accounted for within our cognitive lexical-paradigmatic framework. In the following two chapters, the two pillars of this framework are described in more detail. Chapter 2 focuses on the semantics of causative constructions arguing for a paradigmatic approach. Chapter 3 will present our cognitive view on lexical categories. The two chapters will provide the background assumptions which will be further substantiated in the descriptive analyses in Part II.

CHAPTER 2 THE SEMANTICS OF CAUSATIVE CONSTRUCTIONS

When talking about processes in general and causative processes in particular, linguists usually concentrate on the physical processes and very often on acts of destruction. A probable reason for this predilection is that these verbs express a straightforward kind of causation, in which a human agent directly affects a change of state in a secondary participant, the patient. As indicated by Lakoff, this type of interaction forms the prototype of causation, which is "built into the grammar of languages'" (1987:55). In other words, verbs such as hit, punch, kill, or break are often foregrounded in linguistic treatises on causation, precisely because of their salient position in the conception of causation. In this chapter I will clarify my position vis-à-vis the linguistic tradition concerning the coding of processes and causation. In the literature, the discussion of lexical causatives has often included references to analytical causatives. While these are not immediately relevant to the descriptive analyses in this book, it is warranted that I briefly consider the differences between analytical and lexical causatives, as an introduction to the more elaborate analysis of the latter. Hence, in the first section of this chapter, I will take issue with the often-made claim that lexical and analytical causatives are semantically equivalent (cf. McCawley 1968, Dowty 1972, Hoekstra 1992) and demonstrate that it is linguistically inadequate and too much based on a universalist interpretation of extra-linguistic events. Even studies which tried to counter the alleged synonymy of lexical and analytical causatives (e.g. Fodor 1970, Katz 1970) are often preoccupied with the ontological situation. In their objections to the generative semanticists, most of these researchers draw on the criterion of the spatio-temporal unity of cause and effect that is supposedly implied by lexical but not by analytical causatives. This account runs into problems when it appears that certain lexical causatives can refer to a situation where cause and effect are not contiguous in time and/or space. In other words, both the defenders and the refuters of the synonymy between lexical and analytical causatives falsely assume, as DeLancey correctly observes, that "anything and everything that can be said about a situation is part of the semantic representation of any sentence which describes that situation" (1991:345). I will argue that lexical and analytical causatives are semantically nonequivalent because they encode different conceptualizations of a given situation.

22


Similarly, construal differences within the class of lexical causatives are to be regarded as reflecting differences in conceptualization. Evidence unequivocally shows that there are two types of lexical causatives. But what exactly is the conceptual difference between these two types of lexical causatives? Following Davidse (1991, 1992), I will argue that the difference reflects the distinct transitive and ergative systems. I will trace the increasing sensitivity in the linguistic tradition to the importance of transitivity and ergativity in the grammar of English. My own position synthesizes pertinent insights from predominantly Cognitive Grammar and Systemic-Functional Grammar (next to other sources). This cognitive lexical-paradigmatic approach is only briefly described in this chapter (section 2.2.5) as it will be amply illustrated in the descriptive analyses (Part II) which show how this approach adds a new and revealing dimension by exploring how the more general grammatical principles are influenced by the semantic structure of the verbs that instantiate them. 2.1 Analytical versus lexical causatives In various descriptions of causation, the verb kill has received prominent attention especially in view of its often defended, yet equally often refuted, synonymy relation to cause to die. Since it is also the item central to my data analysis, the following discussion will take kill as the starting point. According to Wierzbicka (1980) the characterization of kill in terms of "cause to die" has a die-hard tradition (she traces it back to the 17th century grammarian Wilkins) but, undoubtedly, the most well-known defenders of this characterization were the generative semanticists. Their analysis of kill postulated a decomposition into semantic primitives: they argued that kill derived (via lexicalization) from an underlying structure "cause to die" or "cause to become not alive". The same deep semantic structure was assumed to underlie cause to die, which accounted for the synonymy between the two surface expressions (see in this respect McCawley 1968 and Lakoff 1970).10 The generative analysis of kill in terms of [Cause[Become[Not Alive]]] has received severe criticism. The most wellknown reactions are probably Fodor (1970), Katz (1970), Shibatani (1972, 1973, 1976) and Wierzbicka (1980). These reactions counter McCawley's analysis, yet are still inadequate in their account of the complex meaning structure of kill. Furthermore, while all of these studies try hard to pin down the difference between lexical causatives (like kill) and analytical causatives (like cause to die), they overlook or do not account for the two-fold distinction that characterizes the class of lexical causatives itself, represented, for example, by 10

In this generative analysis, the 'cause' in the deep structure is an abstract verb which, although part of the lexical deep structure of the English verb cause, is not to be equated with it (cf. Dowry 1972).

THE SEMANTICS OF CAUSATIVE CONSTRUCTIONS

23

kill and break. In this section, I will reproduce some of Fodor's, Katz's and Wierzbicka's criticisms and discuss the remaining inaccuracies in their positions. Fodor (1970) gives the following reasons for arguing against deriving both the lexical causative kill and the analytical cause to die from an underlying composite structure "cause to die". First, if both "cause Mary to die" and "Mary die" are constituents in the deep structure of John killed Mary, then it should be possible, Fodor says, to have an anaphoric do so-construction in which " 'do so' replaces the VP in the constituent sentence 'Mary die' " (1970:431). But, as his examples clearly show, while the do so-construction is possible with the analytical causative, it is impossible with kill11: (l)a. John caused Mary to die and it surprised me that he did so. (Fodor 1970) b. John caused Mary to die and it surprised me that she did so. (2)a. John killed Mary and it surprised me that he did so. (Fodor 1970) b. *John killed Mary and it surprised me that she did so. Secondly, Fodor observes that in the case of the analytical cause toconstruction, "one can cause an event by doing something at a time which is distinct from the time of the event" (1970:433), whereas such a separation in time and/or space is not possible for lexical causatives: (3)a. John caused Bill to die on Sunday by stabbing him on Saturday. (Fodor 1970.) b. * John killed Bill on Sunday by stabbing him on Saturday. On the basis of this, Fodor characterizes analytical causatives as depicting 'compound' and lexical causatives as depicting 'atomic' causal events. Although there is a lot to be said for Fodor's distinction between 'atomic' and 'compound' events, it appears that Fodor is referring to referential time, i.e. the temporal sequencing as it unfolds in the world. If this is so, his account fails to explain why it is nevertheless in principle possible, and unproblematic, for kill to be used in cases where referentially the "kill-component" is separated from the "die"-component. Let us consider some examples from our corpora. The first example is an excerpt from Ustinov's play The Unknown Soldier and his Wife (LDC), a conversation between General Maximus (G) and his Archbishop (A): G: A:

11

Aah! (He runs the archbishop through with his sword). (Falling to his knees with an echo of the general's shout) Aaah!.. .Thank you, Maximus...

For reasons of clarity I have emphasized the relevant participants that figure in the individual clauses.

24


G: A: G: A: G: A: G:

(Aghast) What have you made me do? I have taught you remorse...I have given you a conscience as a parting gift... Why did you make me kill you? Because you are weak, my son. There's no weakness more pernicious than the weakness of the strong. (Examining the archbishop as though he were some curious, half-dangerous animal) Why are you smiling? (Begins to laugh very softly, and shakes his head feebly) What a question. (Slow) Do you derive pleasure from suffering? (No reply. The general loses his temper. He stabs the archbishop again and again.) There! There! Enjoy yourself! (The archbishop rolls over and dies. The general stands there for a moment, panting. He returns his sword to its scabbard.)

Although the use of kill is perhaps somewhat peculiar (but well-suited for dramatic purposes), it is still quite plausible, despite the temporal separation between the first lethal thrust of the General's sword and the Archbishop's dying (admittedly, the time span is considerably less than that in Fodor's example, but it is nevertheless irrefutably present). This example shows that something more than mere 'time matching' is involved. More specifically, it is the conceptual integration of the two components rather than their referential simultaneity which determines the acceptability. The General's use of kill is appropriate as it incorporates the subsequent dying of the Archbishop as part of his kill-action, i.e. his action (the thrusting of his sword) and the Archbishop's dying are conceptually tied together as one event, regardless of whether the actual realization in the world 'out there' might involve some temporal sequencing. In any case, the focus is on the General's deed, as is also indicated by his questions What have you made me do?... Why did you make me kill you? When you come to think of it, the use of kill in Ustinov's play may not be so unusual after all, since temporal sequencing is almost inevitably present in any kill-event (coded by kill or another verb of killing). The most unmarked construal of caused events is one in which the cause precedes the event, and often the two are known not to fully coincide in time.12 For instance, in the sentence (4)

They went straight from trains to "showers" under Zyklon B, which killed in minutes... (WSJ)

12 A construal in which the result precedes the cause, which Talmy (1985b) has termed "resulting-event causation" is illustrated by sentences like The vase broke from a ball's rolling into it or He died from a car hitting him (Talmy's examples).


25

the adverbial in minutes emphasizes the efficiency of the poison gas, yet also acknowledges that there is some time lapse, a few minutes at most, between the emission of the gas and the lethal result intended. The time lapse does not, however, affect the acceptability of kill. Finally, consider the following example which I heard on the radio. On March 23, 1994, the Mexican presidential candidate, Luis Colosio, was assassinated at a meeting in Tijuana. The following morning, the newsreport on the Flemish radio station (BRTN) opened by saying that the presidential candidate was "shot dead" (doodgeschoten) and then continued to specify that two gunmen had shot him in the head and chest. The mortally wounded victim was brought to the hospital, where he died of his injuries (for which the broadcaster used the expression gestorven, the past participle of sterven "die"). If we take Fodor's observations as pertaining to referential time, the use of doodgeschoten would have been problematic since the man was not shot dead but mortally wounded and died later in the hospital. Yet, the use of doodgeschoten is quite natural and even well-suited to open the newsreport because it directly specifies the essence of the matter. At the risk of arguing the obvious, I maintain that lexical causatives such as kill almost invariably code events in which the agent's deed and the patient's affectedness fail to coincide referentially. However, unlike analytical causatives (like the cause to-construction), lexical causatives neutralize this difference and code the components as concurrent and integrated. If the examples above felicitously employ kill to code events in which the cause and effect do not coincide in referential time, why then is such integration impossible for Fodor's example *John killed Bill on Sunday by stabbing him on Saturday? The sentence is unacceptable because the temporal framing is related to points of reference that are too explicitly differentiated, viz. Saturday and Sunday. Fodor's observation may hold for the particular example, but he misses the point by interpreting the "adverb matching" referentially, as do many descriptions of causation. Another line of criticism against McCawley's analysis is offered by Katz, who criticizes it as "a false account of the meaning of the formative 'kill' ", (1970:253) since it does not make a distinction between 'direct' and 'indirect' causation. Katz is probably the first to counter the "cause to die" analysis on the basis of this distinction, which since then has figured prominently in the discussion of lexical and analytical causatives (e.g. King 1988; Lakoff 1987; Lakoff & Johnson 1980; Shibatani 1972, 1973; Smith 1978; Vendler 1984; etc.). Drawing on a Wild West anecdote, Katz argues that someone (the gunsmith) may be indirectly responsible for someone else's death (the sheriff's, shot by some outlaw), but that this does not entail that this gunsmith killed the sheriff. This line of criticism has affinities with Fodor's observation about the

26


possible spatio-temporal distance between the causation and the caused event in analytical causatives, since the directness of causation is in a way also a kind of distance, either referential or conceptual. In other words, the more remote the agent's deed is from the effect (remote in whatever sense), the more contextual support is needed to justify the mental gymnastics of integration into a lexical causative. As Lakoff (1987:55) phrased it, "the more direct the causation, the closer the morphemes expressing the cause and the result5' (cf. also Lakoff & Johnson 1980:Ch.20; Haiman 1983). However, this iconicity principle is not entirely accurate as it relies on the referential situation, whereas what is at issue is a difference of conceptualization of possibly one and the same reality. Sentences like (5) (6) (7)

John killed Bill. John caused Bill to die. Bill passed away.

may all refer to the same event, yet differ in the way this event is conceptualized. A case in point is the following example: (8)

I protested against Hitler's anti-Jew laws. He killed off six million of us! (LDC).

in which Hitler is coded as the agent directly responsible, despite the fact that in reality he was quite remote to the actual killing. Against the mainstream interpretation, I will argue, following Davidse (1991), that the iconicity principle has validity only if interpreted non-referentially, i.e. when the conceptualization is iconic for the semantics of the participant-process relation in the grammatical construal (cf. below). Another problem of the iconicity principle is that while it holds for kill, it does not hold for all lexical causatives. In line with Fodor's criticism, Wierzbicka (1980:162) argues that "a semantic component like 'somewhere' " is part and parcel of the semantic structure of a sentence like Peter killed the cat in the attic, because "both killer and victim are always thought of as being in one place". Unity of place is not implied by Peter caused the cat to die in the attic. Thus, she finds in the meaning of kill "two simultaneous events, one involving the doer and one involving the victim" (1980:165), whereas cause to die allows for a separation of doer and victim. Attempting to shed light on the issue, she brings in other verbs, such as harden (as in John hardened the metal), solidify (as in John solidified the metal), and thicken (as in John thickened the sauce). In each case, she claims that two simultaneous events can be found: John is doing something and, at the same time, the metal is becoming harder or the sauce is becoming thicker. Her solution parallels Fodor's analysis of the verb


27

melt: "one can cause an event by doing something at a time which is distinct from the time of the event. But if you melt something, then you melt it when it melts" (1970:433) and this, he maintains, also applies, mutatis mutandis, to kill. Fodor and Wierzbicka are both correct in pointing out that the conceived spatio-temporal co-extension of the action and the (coming about of the) effect are central to the meaning of kill, but both incorrectly level it with verbs such as melt or solidify, which differ from kill in allowing a spatio-temporal separation, as already observed by Lakoff and Ross (1972). In their discussion of anaphoric island phenomena, they observe that for some causative verbs the spatiotemporal cohesion is not compulsory. Consider the by now famous examples (cf. also Shibatani 1973:289): (9)a. The scientist finally hardened the metal, but it took him a year to bring it about. b. *John finally killed Mary, but it took him six months to bring it about. Lakoff and Ross try to salvage the analysis of both verbs as causatives by postulating a principle that constrains anaphoric reference to cases where the embedded intransitive verb is morphologically related to the transitive verb. This is true for hardenintr/hardentr, but not for kill/die, and hence the unacceptability of (9b). The ungrammaticality, they conclude, "cannot be taken as a counterargument to the causative analysis, contrary to Fodor's claim" (Lakoff & Ross 1972:125). I agree with Davidse (1991) that the morphological constraint postulated by Lakoff and Ross does not really explain why 'a-type' verbs allow the non-coextension whereas 'b-type' verbs do not. The above discussion has clearly shown that lexical and analytical causatives cannot be put on a par, but encode a different conceptualization. As our main concern is with lexical causatives, the analytical causative constructions will fairly much be ignored in the present work (see Davidse 1991:65-77 for some further discussion). In the course of the discussion it has been noted that not all lexical causatives behave in the same way, as illustrated by the differences between kill and melt. Is this difference significant? Most of the cited studies tend to ignore the difference, yet it has not gone by unnoticed in (more) recent studies. Langacker (1991b:408) observes that there are of course many verbs—such as kill, throw, and chop—in which the notion of causation is inherent yet sublexical, in the sense of not being coded individually. Though still implicit, that notion is somewhat more 'visible' to the analyst with verbs like break, open, and cook, for it constitutes the only semantic difference between their intransitive and transitive variants.

28


Langacker makes a distinction between verbs in which the cause is sublexical and those for which it stands out more. This distinction strikes at the heart of the contrast between the transitive and ergative paradigms operative in the English grammar. Since the two paradigms are pivotal to the analysis presented in this book, it is essential that they are more closely examined. 2.2 Transitive versus ergative lexical causatives The early generativists' studies, as well as their critiques discussed above, neglect the distinction between transitivity and ergativity. Ironically, around the same time as McCawley's first analysis of kill in terms of "cause to die", M.A.K. Halliday published a more refined view on transitivity in his three part exposition "Notes on Transitivity and Theme in English 1-3" (1967-8). In these articles, he proposes a considerably more refined analysis of related one and two-participant construals in English, distinguishing between what at that stage he referred to as the "effective" and the "descriptive" systems. Davidse (1991, 1992) builds on Halliday's insights and interprets the effective and descriptive systems as paradigmatically—not morphologically—encoding the transitive and ergative principles that govern the grammar of actions and events. Since I will elaborate her insights towards the end of this discussion, I will refrain from going into Halliday's views here. Instead, I will start with a short description of transitivity and ergativity as used in traditional typology which will provide the background for the semantic reinterpretation offered in both Langacker's cognitive and Davidse's paradigmatic approach. 2.2.1 Traditional form-based transitive/ergative typology The traditional views on ergativity come from typological studies and are based on morphology. Ergativity is restricted to languages with ergative case marking or syntactic patterning, such as 'ergative' co-ordination patterns in complex sentences in (some) Australian Aboriginal languages (see e.g. Dixon 1979, 1994, Plank 1979, and Comrie 1973, 1978, 1981 for descriptions within the traditional typology). A language is ergative if "the subject of an intransitive clause is treated in the same way as the object of a transitive clause, and differently from the transitive subject" (Dixon 1994:1). In contrast, in a nominative/accusative system, the subject of transitives and intransitives are coded alike (nominative), whereas the transitive object is marked differently (accusative). The ergative system is illustrated by Dixon's examples from Dyirbal, examples (10) and (11); the transitive, by his examples from Latin, (12) and (13). (10)

numa banaga-nyu. (Dixon 1994:10) father-ABS return-NONFUT "Father (S) returned"

THE SEMANTICS OF CAUSATIVE CONSTRUCTIONS (11) (12) (13)

29

numa yabu-ngu bura-n. (Dixon 1994:10) father-ABS mother-ERG see-NONFUT "Mother (A) saw father (O)" dominus venit. (Dixon 1994:9) master-NOM come-3RDPERS-SG-PRES "the master (S) comes" dominus. servum audit. (Dixon 1994:9) master-NOM slave-ACC hear-3RDPERS-SG-PRES "the master (A) hears the slave (O)"

Dixon diagrams the relations for a nominative/accusative system and for an absolutive/ergative system as in Figure 2. NOM/ACC

ERG/ABS A

ERGATIVE

LEGEND

ABSOLUTIVE

S intransitive subject A transitive subject 0 transitive object

NOMINATIVE

_ S ACCUSATIVE

O_

Fig. 2: Dixon's representation of nominative/ergative typology I will not elaborate on Dixon's typology, but take it as reflecting the mainstream view on transitivity/ergativity. There are, however, two important comments on Dixon's views: the first is a methodological objection, the second relates to the descriptive inadequacy of his observations. The methodological objection concerns Dixon's idea that a category can be set up only if there is a reflex in form, which in this case should be either morphological marking or overt syntactic patterning in one syntagm. However, categories are also formally realized in different ways than these two allowed by Dixon, viz. by the systematic relations linking all the paradigmatic variants of a construal type. This methodological point has been made by researchers like Whorf (1956) and more recently by Levin (1993). Whorf (1956:87-101) distinguishes overt grammatical categories, or phenotypes, from covert grammatical categories, or cryptotypes. Phenotypes find their realization via a specific formal marking or syntactic patterning. Cryptotypes make themselves felt via their correlation with patterns of syntactic behaviour (cf. also Deane & Wheeler 1984). In Whorfian terms, Dixon's typology can thus be criticized as focusing exclusively on the phenotypical level of clause grammar and neglecting the cryptotypical level. Cryptotypes are difficult to define and can only be arrived at by examining their different exponents systematically. Only through systematically tracing these exponents, viz. the

30


various interactions with other patterns in the language, can one arrive at a proper description of the realization and semantic value of 'crypto-grammar'. Levin (1993) can perhaps be regarded as the first 'mainstream' linguist who systematically 'twists and wrenches' verbal constructions. As such, she replaces all the ad hoc arguments used in the literature by a systematic descriptive tool, viz. that of verb alternations. She shows that alternations between construals are a crucial heuristic to the semantics of verb meaning (see section 2.2.3 for a critical discussion of Levin's approach). The second point of criticism levelled at Dixon's description is its inadequacy for English. In his 1979 article, he acknowledged that English has "lexical ergativity", but rejects this as inconsequential because it is not based on overt formal realizations. He further conflates transitive and ergative oneparticipant verbs (e.g. die vs. break), a distinction which will be demonstrated to be crucial in the English grammar of processes. In his recent book on ergativity, Dixon openly objects to the use of the term ergativity for English (pertaining to 'ambitransitive' verbs like break, burst etc.), as in his view it is indisputably an accusative language. He says that "[t]he use of 'ergative' in this lexical-semantic/causative sense is quite illicit (in addition to being potentially most confusing) in terms of the well-defined use of the term." (Dixon 1994:20). Dixon's con&sion results, once again, from his exclusively formal view on ergativity. Following Davidse (1991, 1992), this work reveals that ergativity is important to English clause grammar and manifests itself in a variety of ways. Against the background of the above discussion, the following sections introduce the vast complexities of the transitive and the ergative systems operative in the English grammar of actions and events. The description is organized around four sections. The first one presents Langacker's conceptual explanation for the nominative/accusative and ergative/absolutive systems. The second section takes the distinction to a more systematic level, discussing Levin's (1993) categorization of verbs based on the type of alternations they allow (although Levin herself does not distinguish between ergative and transitive). The third section discusses Davidse's (1991) research. She synthesizes many of the relevant insights from other studies, but she regards the issue of transitivity versus ergativity as one of clause grammar and the semantics of construal types. The last section presents my own lexical-paradigmatic position as a further refinement to Davidse's framework, drawing on insights from my analysis of verbs of killing. 2.2.2 Langacker's cognitive approach Langacker largely follows the traditional typological definition of the nominative/accusative and ergative/absolutive systems, He defines the former (henceforth called the 'transitive' system) as one in which "intransitive and


31

transitive subjects have the same form, and the direct object is marked differently" (199lb:378). In contrast, an ergative/absolutive system (henceforth, the 'ergative' system) is one in which "a transitive subject is specially marked, whereas an intransitive subject has the same form as a transitive object" (1991b:379). For languages with case marking, we thus have the following schematization (taken over from Davidse 1991:25): NOM/ACC

ERG/ABS

verb verb

ACC

Fig. 3: Typology of nominative and ergative systems Langacker departs from mainstream work on case systems by characterizing case markings as inherently meaningful, Moreover, the transitive and ergative patterns are not only coded by morphological markings, but find "numerous other linguistic manifestations" (1991b:381). In fact, "every language probably uses both patterns in one fashion or another, though the mixture varies and a particular pattern is often predominant" (1991b:381).13 From this it follows that each pattern has a cognitive motivation. But what precisely is the conceptual basis for the transitive/ergative distinction? Langacker's description assumes the existence of a cognitive model directly relevant to our conception of actions and events, the 'billiard-ball model'. This archetypal folk model presents our world as being populated by discrete physical objects which, driven by or transmitting energy, are capable of moving through space and making contact with one another (cf. also Talmy 1985a for a similar view in terms of force-dynamics). Transmission of energy may cause the affected object to move or change and possibly to interact with other entities. Such interaction can be represented as an action chain, illustrated as in Figure 4 below (cf. Langacker 1991b:283). The double arrow symbolizes transmission of energy; the change of state which the given object undergoes is represented by the squiggly arrow.

13

Similarly, Halliday observes that "probably all transitivity systems, in all languages, are some blend of these two semantic models of processes, the transitive and the ergative" (1985:149). Also Dixon (1994) acknowledges that languages mostly show characteristics of both.

32


Fig. 4: The action chain Our experience of interacting entities gives rise to yet another model, namely, the 'stage model'. This model "idealizes an essential aspect of our ongoing experience: the observation of sequences of external events, each involving the interaction of participants within a setting" (Langacker 1991a:210). A sentence like Jill killed Jean with a baseball bat implicitly incorporates the notion of an external observer who observes Jill as energetically acting upon the bat which in turn acts upon Jean. From our experience of such events, we develop idealized conceptions of typical roles that event participants can assume. In the example sentence, Jill fulfils the role of Agent, the baseball bat that of Instrument and Jean that of Patient. Langacker argues that these role archetypes are "pre-linguistic conceptions grounded in everyday experience" (1991b:285). The two models combine to form a more complex conceptualization of events, the 'canonical event model', which can be diagrammed as in Figure 5 (1991b:285).

Fig. 5: Canonical event model Obviously, the canonical event model is a generalization that abstracts from events as coded by particular verbs. As Langacker states it, "at the extreme, every verb defines a distinct set of participant roles that reflect its own unique semantic properties" (1991b:284). Let us return now to the difference between the transitive and the ergative systems, which Langacker characterizes in terms of the distinct construal paths. The participants marked NOM and ABS (or their equivalents in languages without overt case marking) are usually zero-marked, which indicates that they are the conceptual building blocks of the transitive and ergative systems respectively, as is also indicated in Figure 3. The natural construal model for the transitive system is the action chain. The participant marked NOM is head of the


33

energy flow and thereby has the status of primary clausal figure. The participant marked ACC lies downstream and is the secondary clausal figure. Thus, NOM is more basic as it marks the participant that is the starting point of the energy flow. By contrast, in an ergative system, the participant that is conceptually basic (marked ABS) is the secondary clausal figure, i.e. the one that lies downstream from the energy source. Consequently, the conceptual path runs counter to the flow of energy along an action chain. This path is defined in terms of conceptual autonomy/dependence (henceforth A/D) inherent in the structure of conceived events. More specifically, the A/D layering means that in a complex event conception "it may be possible to conceptualize certain event components independently of the others, which makes them conceptually autonomous" (1991b:286). More specifically, it is thus in principle possible, "even for events that are saliently energetic [...] to factor out the force-dynamic component and conceive of the change of state autonomously" (1991b:289). Talmy (1985b) expresses similar views, saying about the sentence the vase broke that it codes an "autonomous event", in the sense that "[it] presents an event as occurring in and of itself, without implying that there is a cause (such causes as there may be fall outside of attention)" (1985b:79). Figure 6 is an adaptation of how Langacker represents the difference between The glass broke, the hammer broke the glass and Floyd broke the glass. The difference concerns first, the profile that is imposed onto the complex event (represented in bold face) and second, the natural path in which the event is assembled (represented by the leftward arrows underneath each box). (a) The glass broke

(b) The hammer broke the glass

7

(E1(T))

(c) Floyd broke the glass with a hammer

(E2(E1(T)))

Fig. 6: Scope, profile and A/D layering of a complex event conception Working outward from the autonomous nucleus, the assembly of a complex event conception implies a mental path which starts with the 'theme' (T) and moves 'upward' towards the energy source (E) of the action chain. This results in event conceptions of progressively greater complexity, represented by Langacker (1991b:382) as [T] > [E1[T]] > [E 2 E 1 T]]]. The thematic participant is the participant that is evoked as part of the conceptually autonomous core of a processual predication. In the sentences represented in Figure 6, the thematic

34


participant is the glass. The diagram nicely visualizes that the contrast between construal type (b) and (a), which has become known as the causative/noncausative (or inchoative) alternation14, represents a difference of conceptualizations of the event and the two construals are thus semantically non-equivalent. Langacker observes that there are many English verbs, e.g. break, open, melt, starve, burn, freeze, etc. "that can be used either transitively or intransitively without any difference in form" (1991b:387). These verbs, says Langacker, display an ergative pattern. However, Langacker fails to point out that factoring out the causation component is not always possible. Verbs such as kill, murder, execute, slay, hit, cut, etc. do not allow the non-causative construction. These verbs do allow a formally 'intransitive' construction, as in soldiers going to war must be ready to kill, but this sentence does not mean that the soldiers must prepare to die, but that they must be prepared to "be the cause of somebody else's dying". While Langacker's characterization of the ergative model in English is partially correct, he offends against his own criterion when he includes in the set of ergative patterns a construction like the tomato peels easily. In his opinion, this construction, which is generally called a 'middle' or 'medio-passive' construction, "requires its theme to have a skin-like covering (It peeled easily; He peeled it easily)" (199lb:388). However, the tomato is not a conceptually autonomous thematic participant as it cannot occur in a non-causative construction: compare *the tomato peeled with the tomato burst. A descriptively more adequate analysis of the middle construction is postponed until Chapter 4, section 4.2. While I do not want to deny the inspirational value of Langacker's description, it does not take the empirical observations to their logical conclusion, which would result in a more systematic treatment of the different patterns. The major inaccuracy of Langacker's analysis arises from his notion of "thematic participant", the participant that stands in a thematic relationship to the process. His characterization of such a thematic relationship as a "conceptually autonomous relationship involving just a single participant" (199ib:286), leads him to incorrectly identify the single participant in a middle construction as thematic. By characterizing the subject of 'true' intransitives like the tree fell as a thematic participant as well, Langacker conflates under one general denominator grammatical relationships that are clearly semantically distinct. In the case of the following sentences

14

I will suggest a different terminology in section 2.2.4 which will reflect more closely the paradigmatic oppositions at issue. Until then, I will continue to use the traditional terms 'causative/non-causative'.


(14) (15) (16)

35

The tree is falling. The window opened. The tomato peels easily.

Langacker attributes to the subject the role of thematic participant and he identifies the construal as 'absolute', i.e. when "the autonomous event component is evoked or profiled independently" (1991b:389). His notion of absolute construal is too schematic in its alignment of (14) and (15), since for (15) the cause can be incorporated into the clause, e.g. John opened the window, whereas this is not possible for (14), e.g. *John is falling the tree. I argue that this account violates the more subtle distinction in English clause grammar. From Langacker's account it follows logically that something that is conceptually dependent cannot be isolated in an absolute construal, as witnessed by his examples. (17)a. The wind caused the tree to fall over. b. The tree fell over. c. *The wind caused. The acceptability of the absolute construal in (17b) is explained by the fact that "we can [...] readily conceptualize this change as occurring autonomously, i.e. in the absence of any external force" (1991b:287). The infelicity of (17c) is due to the fact that "causation is conceptually dependent—the very notion makes inherent (albeit schematic) reference to the change induced" (1991b:287). We will not deny that in many cases an inherent reference to the caused process is present, yet this does not prevent well-formed one-participant structures similar to (17c) to occur. Consider some examples of objectless transitives: (18) (19)

These are soldiers trained to kill. Bill always interrupts. (Rice 1988)

The (unfortunate) logical conclusion of Langacker's analysis is that the soldiers and Bill are not in a thematic relationship with the process, as he says this relation to be restricted to the subjects of (1) 'true' intransitives (the tree fell), (2) non-causatives (the window opened), and (in my view incorrectly so) (3) middle constructions (the tomato peeled easily). Yet intuitively (and we like to believe that Langacker himself would subscribe to this intuition), the oneparticipant constructions with kill and interrupt are felt to profile what is conceptually basic, focusing on the agent's action. Non-causatives, in contrast, focus on the patient (cf. below).

36


To summarize, Langacker's account clearly needs a more systematic and allencompassing analysis which does justice to the subtle semantic distinctions operative in English clause grammar. 2.2.3 Levin 's lexical-semantic approach In her most inspiring work on English verb classes and what she identifies as 'diathesis alternations', Levin (1993) goes a long way towards explaining, in a systematic fashion, the "alternations in the expression of arguments, sometimes accompanied by changes of meaning" (1993:2), in which the English verbs may participate. She is committed to the idea—and I believe correctly so—that speakers of English know which diathesis alternations a verb can have, but also notes that this goes beyond "the type of lexical knowledge traditionally represented in subcategorization frames" (1993:2) The speakers' ability to make judgements about a verb's syntactic behaviour leads Levin to the assumption that "general principles of grammar are at work, allowing the syntactic behaviour of a verb to be predicted from its meaning" (1993:5). She presents her systematic study of alternations as a heuristic tool towards verb meaning and towards a general classification of verbs. The classification which is most relevant to the present purpose is the distinction between the BREAK verbs and the CUT verbs. The BREAK verbs

denote "an entity undergoing a change of state" (1993:9), and allow the noncausative construction (e.g. The cup broke). The CUT verbs, in contrast, do not allow it. The distinction that emerges is thus the following: (20)a. b. (21)a. b.

The window broke. (Levin 1993) The little boy broke the window. Margaret cut the string. (Levin 1993) *The string cut. (on the interpretation "became cut")

Trying to explain the causative/non-causative alternation with BREAK verbs on semantic grounds, Levin claims that it "is found only with verbs of pure change of state" (1993:10; my emph.). Although Levin's explanation is plausible and applies to a large number of verbs, the notion of 'pure change of state verbs' is too general for the delineation of the categories.15 First, the addition of the qualification 'pure' seems to be drawn into the discussion to remedy the overly general notion of 'change of state'. The BREAK verbs, which comprise verbs such as break, crack, smash, splinter, or tear, are representative examples of the pure change of state verbs allowing the non15

Fellbaum (1988:129) quotes Hale & Keyser (1987) as relying on a similar criterion for the class of BREAK verbs and they suggest the rule [x cause [y undergo change], (by...)]. She also points out that this rule is "a necessary, but not a sufficient condition".


37

causative alternation. By contrast, verbs such as chip, slash, mash, pulverize or squash, which Levin categorizes as CUT verbs, could then be expected to be 'impure' change of state verbs, since they do not allow it. Yet this group undeniably incorporates the notion of a change of state, so much so that in a subsequent categorization Levin herself (1993:241) incorrectly includes chip (a CUT verb) with the BREAK verbs. In general, the CUT verbs do incorporate a more pronounced reference to an agent that causes the change of state, but that does not make them any less change of state verbs. In short, Levin's 'change of state' criterion is too broad to capture the necessary distinctions. Secondly, the 'change of state' criterion is linguistically inadequate, since it fails to give an account of why the causative/non-causative alternation is not always applicable to the same verb in different environments (cf. also Davidse 1991:63). For instance, in Levin's classification, the verbs change and open are 'pure change of state' verbs allowing the non-causative construction (e.g. John opens the shop at 9:00 a.m. vs. The shop opens at 9:00 a.m.). However, for some clauses with these verbs, there is no corresponding non-causative: (22)a. b. (23)a. b.

John opened a tin of baked beans. (Davidse 1991:63) *A tin of baked beans opened. Mother changed the baby. (Davidse 1991:63) *The baby changed (on the interpretation "the baby is clean").

The constructional variability with the same verb is one of the main issues addressed in this book; I will return to it later in this Chapter, in sections 2.2.4 and 2.2.5. In the final conclusions to this book, the issue will be reconsidered in more theoretical terms. To sum up, Levin's 'change of state' criterion fails in two ways: first, it leads to an incorrect classification of verbs and secondly, it ignores the constructional variability that may occur with one and the same verb. Moreover, like Langacker, Levin overlooks the complexity of the middle construction. The middle construction, says Levin, is possible with both BREAK and CUT verbs: (24) (25)

Stale bread cuts easily. (Levin 1993) Crystal vases break easily. (Levin 1993)

Once again, this characterization is too general and must be redefined in view of the semantics of both the construction and the verb. As already said, the middle construction is dealt with in Chapter 4, section 4.2, in which I present additional support of my critique on the change-of-state criterion. There are two additional points of criticism regarding Levin's theoretical assumptions. First, Levin's basic hypothesis is that a verb's syntactic behaviour

38


is predicted by its meaning. While she acknowledges that this hypothesis is not uncontroversial and may eventually turn out to meet with limited success, she maintains the hypothesis, provided, she adds, that one identifies "the appropriate representation of verb meaning". Ironically, her own identification of verb meaning that supposedly conditions the difference between CUT and BREAK verbs leads to incorrect predictions and analyses. The meaning of a verb is no doubt pivotal to a proper description of its constructional behaviour yet Levin overrates the extent to which lexical verb meaning determines this behaviour. She relies too much on the semantic properties of verbs, which she sees as independent of the constructions in which they occur (cf. also Goldberg 1995 for similar criticism). For example, objectless constructions like Soldiers trained to kill or You drink too much cannot be explained solely on the basis of the semantics of the verb. In Chapter 5, section 5.4, I argue, amending Rice's (1988) lexically-based analysis, that these constructions can be characterized properly only if we take into account the meaning of the grammatical construction itself and the way it 'modulates' the meaning of the verb. Secondly, and this follows logically from the previous point, Levin fails to see that the constructions in which verbs occur are meaningful in and by themselves. Her overgeneralization with respect to the middle construction is due to insufficient attention to the "constructional polysemy" of this construction (the term 'constructional polysemy' is Goldberg's 1992, 1995). In the introduction to her work, Levin says that verb alternations are "sometimes accompanied by changes of meaning" (1993:2). Against this view, I posit that alternations are inevitably accompanied by changes of meaning since they reflect different ways of construing an event. While Levin sees meaning as independent of clause structure, I hold the view that the clause is where verbal and constructional meaning meet to form a semantic complex capable of triggering unpredictable (yet motivated) creative usages. Levin does not regard such construal variations. This explains why, somewhat surprisingly given her concern with verbal semantics, her distinctions fail to account for cases like (22) and (23), in which verbs that typically belong to a category which allows a certain alternation occur in contexts for which the given alternation no longer applies. In later work, e.g. Levin & Rappaport (1995), a more nuanced view is defended which in many ways come closer to what the present work argues for. A more critical evaluation of Levin & Rappaport (1995) is given in the final conclusion to this work. To conclude this brief description, Levin's work can be said to provide an impressive overview of verb classes for which she uses the heuristic tool of a verb's alternations. This allows her to set up a general categorization of verb types. However, her work falls short of recognizing some fundamental general patterns that govern (certain aspects of) English clause grammar. At the same


39

time, she adopts a view on verbal semantics that is too absolute and she consequently overlooks the verbs' polysemy which may trigger changes in the type of alternations allowed. The following two sections will address both shortcomings. First, I will consider the general transitive and ergative paradigms at work in the grammar of material processes (section 2.2.4) as described by Davidse (1991, 1992). In the last section (2.2.5), I will characterize how her model should be refined by bringing the semantic flexibility of verbs into the discussion, which is essential in a proper description of English clause grammar. 2.2.4 Davidse's paradigmatic approach Davidse (1991) characterizes the grammar of material processes as "Janusheaded", i.e. governed by two distinct systems, the ergative and the transitive models of causality.16 These models determine not only the constructional potential of verbs, but also verb-derived grammatical structures, like nominalization or affixation. I adopt Davidse's dictum that one cannot even begin to describe with any accuracy the grammar of actions and events in English, if one does not take into account the existence of the transitive and ergative systems. But if one does, many long-standing problem-cases in descriptive linguistics can be given a more precise paradigmatic location [...] and can also be satisfactorily characterized and explained. (1991:19-20)

In what follows, I will give a general presentation of the transitive and the ergative paradigms, without elaborately reproducing the various exponents (syntactic and/or morphological) which Davidse provides as evidence in support of the workings of the paradigms. I should further note that my concern is with material processes only ('the grammar of actions and events observed out there'), since the other two large categories of processes which Davidse discusses, i.e. relational processes ('the grammar of semiotic relations'; e.g. symbolize, resemble, etc.) and mental processes ('the grammar of consciousness and its projections'; e.g. please, like, look, worry, etc.) are not directly relevant to my purpose. Davidse offers an original synthesis of a number of studies17 by positing transitivity and ergativity as distinct networks of construals. These two semantic models project different process-participant constellations in their conception of events, i.e. different ways in which the participants participate in the process. Halliday (1967-8) talks about differences in the "inherent voice" relation between the participants and the process. 16

See also Davidse (1992, 1994) for additional discussions. The most important of these include various publications by Halliday, Hopper & Thompson (1980), Keyser & Roeper (1984), Rice (1987a,b), and Langacker (1991b). 17

40


In an ergative two-participant structure, the participant affected by the event is presented as the one most crucially involved; in addition to being affected it also co-participates in the process. Davidse terms this participant the 'Medium', a term that I will take over (It may be helpful to think of a medium in the ESP sense: it is affected yet at the same time the centre through which the process manifests itself.) It is the Medium's co-participation (and not Levin's 'change of state') which distinguishes an ergative from a transitive two-participant construal and determines the semantic difference between the two models of causality encoded. The structural centrality of "the MEDIUM-PROCESS nucleus" is reflected in the fact that it is precisely this nucleus which can be isolated in the non-causative construction, The window broke. The semantic value of the noncausative construction is that it neutralizes whether the process was selfinstigated or instigated by an external Instigator. As Smith (1978) puts it, the construction is positively marked for the features of external control as well as independent activity (cf. also Haspelmath 1993:90). However, the ergative system prototypically opens up to the left to incorporate an Instigator, the "causer" of the action (e.g. The boy broke the window), which resolves the voice vagueness.18 The ergative model is thus one of "cause-&-effect" (Halliday 1985:146) or one of "instigation" (Davidse 1991:61). This entails that Guerssel et al.'s (1985:54) view on the non-causative as "having a single passive participant but no active participant" is essentially incorrect: non-causative constructions profile the Medium, an entity that is affected by the process, yet is at the same time also an agentive participant. Figure 7, taken over from Davidse (1991), offers a schematic representation of the ergative model. In 7(a) the straight lines denote the conceptual dependency relations: the process is conceptually dependent on the Medium, the entity functioning as the figure in the ergative construal (indicated in bold face).19 The Venn diagram indicates that the PROCESS-MEDIUM cluster is semi-autonomous vis-à-vis the instigation. In similar vein, Smith (1978:101) talks about the "relative independence of the activity". Figure 7(b) visualizes the ergative system as a nuclear one with two processual layers, the instigated process and the instigation of the process, which need not be co-extensive in time or space (Davidse 1991:67ff). The arrow on the inner circle indicates that the process can be self-instigated.

This term does not cover the same ground as Fillmore's use of it; he uses it in a definition of the agentive case, reserved for "the typically animate perceived instigator of the action" (1968:24). 19 The notion of 'figure' (indicated by bold face) is a cognitive grammar reinterpretation of Davidse's figures.


41

Fig. 7: The ergative paradigm The nuclear ergative pattern is fundamentally different from the linear transitive paradigm which has an inert Affected. The transitive paradigm takes the Agent as its centre. To distinguish this Agent from the ergative Instigator, Davidse uses the term Actor to designate the category. Both the Instigator and the Actor are instantiations of the more schematic category Agent. The basic building block of the transitive paradigm is thus the ACTOR-PROCESS combination, which can, in some cases, be isolated in objectless transitive constructions, e.g. God saves or Soldiers trained to kill (see Chapter 5, section 5.4). In other words, the transitive system is Actor-centred and right-oriented: the event finds its starting point in the Actor, the most central participant, and in the twoparticipant construction it opens up to the right to include the participant onto which the Actor directs his action, the 'Goal'. The Goal does not co-participate in the process, but is an inert participant wholly affected by the Actor's action. As such, the transitive system emphasizes a "deed-&-extension" aspect of a process (Halliday 1985:146; Davidse 1991:23,61). Figure 8 is how Davidse visualizes the difference from the ergative system: in the transitive system, the ACTOR-PROCESS combination is the nuclear building block (a), and the system is a linear one (b) that may extend to the right.

Fig. 8: The transitive paradigm The transitive and ergative paradigms form two distinct models of causality, each imposing its own image on an event. Since they construe a process in a particular way, the models are inherently meaningful and semantically nonequivalent. In addition, they help explain the specific semantic structure of lexical items that instantiate either of them. Obviously, the schemata do not exhaustively define the semantics of the instantiating clause; our research

42


reveals how verbal semantics as well as other elements in the clause (e.g. particles) make a crucial contribution too. Like most conceptual categories, the ergative and transitive paradigms constitute complex networks of related construals, structured around a prototype. A further discussion of their complex internal structure will be postponed until Chapter 3, section 3.2, following a more elaborate discussion of the prototype-based structure of conceptual categories. In line with the schematic representation of the transitive and ergative case typology (see Figure 3), the situation can be represented as follows (taken over from Davidse 1991:25): Transitive

Ergative

Fig. 9: Semantic reinterpretation of the traditional typology Despite Dixon's objections, I subscribe to the view that ergativity is relevant for English, provided that one realizes that it does not manifest itself in overt case markings but in various alternation patterns, which reveal the transitive and ergative models as more covert categories of grammar. As said, the two paradigms make themselves felt in various areas of the grammar. The reader is referred to Davidse (1991) for an elaborate discussion of these exponents; for the present purpose, it suffices to briefly mention some of them as a further clarification. Some of these will be taken up later, when relevant to the analysis of verbs of killing. In view of the above discussion, two-participant constructions are to be distinguished paradigmatically: John killed Mary is a transitive construction, as indicated by the possible alternation with John killed, whereas Bill broke the window is an ergative construction, as indicated by the possible alternation with the window broke. Although formally a one-participant construction, the objectless transitive e.g. John killed is still regarded as a transitive (see Chapter 5, section 5.4). For the one-participant constructions, the paradigmatic opposition is between an ergative construction like the window broke and 'true' intransitives like John died or John is running. The latter are defined as a subtype of the transitive, since they centre around the (prototypically volitional) Actor (see Chapter 4, section 4.3.1, on the relationship between intransitive and ergatives).


43

Process modification ana anaphoric reference, already mentioned in the discussion of analytical causarives, can reveal paradigmatic cuts between a transitive construal (with only one processuai ievei) and an ergative one (with two processuai layers) In a transitive construal, e.g. he flogged the slave for half an hour, the adverbial modifier can pertain only to the ACTOR-PROCESSGOAL unit; in an ergative construal, e.g. John boiled the water for half an hour (Vendler 1984), it may (but need not always) be limited to just the MEDIUMPROCESS nucleus.

The different directionalities of the two paradigms is nicely revealed by reflexive constructions. A transitive reflexive involves the contrast between a non-reflexive and reflexive Goal: John killed himself stands in opposition to John killed his neighbour. An ergative reflexive, in contrast, involves the question as to whether or not the Instigator is co-referential with the Medium: the machine stopped itself is not in opposition to The machine stopped the car, but to Lizzy stopped the machine (cf. also Keyser & Roeper 1984:405 on the all by itself-phrase). Also in morphological processes, different paradigmatic cuts can be discerned, as reflected in, for instance, reprefixation (see Davidse 1991:108ff, reinterpreting observations by Keyser & Roeper 1984), the -able derivation (see Chapter 4, section 4.2.2.3) or the -er suffix deriving agentive nominals from verbs (see Chapter 5, section 5.3). At this point, the traditional terminology can be revised into one that more appropriately captures the different paradigmatic distinctions and commonalities. Davidse uses the term 'effective' for both ergative and transitive twoparticipant constructions and 'middle' to refer to the transitive and ergative single-participant constructions. For the two-participant constructions, I will continue the term 'effective', which is also used in mainstream linguistic descriptions as the semantic counterpart of grammatical transitivity. For instance, Hopper & Thompson say that transitivity is a property of the entire clause, in which the action is "typically EFFECTIVE in some way", as it is " 'carried-over' or 'transferred' from an agent to a patient" (1980:251). By further specifying the paradigmatic opposition, we end up with the contrastive pair 'transitive effective' versus 'ergative effective'. I will not continue Davidse's term 'middle', which I will restrict to the medio-passive construction, e.g. That tent puts up in anybody's yard (Fellbaum 1985), further discussed in Chapter 4, section 4.2. Instead, I suggest 'noneffective' as a cover term for the one-participant constructions and like the effectives, they can be distinguished paradigmaticaliy: what is traditionally called the intransitive (e.g. John died) is thus the 'transitive non-effective' and the non-causative (or inchoative) will be called the 'ergative non-effective'.

44


At first sight this may seem as an unwarranted terminological inflation. However, the terminology proposed here allows a clear presentation on different levels, which can be presented as follows. CONSTRUCTION

TRANSITIVE

ERGATIVE

EFFECTIVE

John killed Mary

John suffocated Mary

[AC-PROCESS-GO]

[IS-PROCESS-ME]

John killed [AC-PROCESS-(GO)]

Mary suffocated

NON-EFFECTIVE

[ME-PROCESS]

Mary died [AC-PROCESS]

PSEUDO-EFFECTIVE

Mary died a gruesome death

He grew a wart (Davidse 1991 )

[AC-PROCESS-RANGE]

[SETTING-PROCESS-ME]

At the most schematic level, the distinction is between effective, non-effective, and pseudo-effective structures. Pseudo-effectives are so called because one participant (underlined in the examples) is not a true participant. As with the effectives and non-effectives, a paradigmatic refinement can be made, since the area of variability is different for both paradigms: for the transitive pseudoeffectives (called Ranged constructions) there is a pseudo-Goal, for the ergative pseudo-effectives (called Setting constructions), a pseudo-Instigator (see Davidse 1991:115-40). The pseudo-effective constructions are not immediately relevant to the field of killing in present-day English, except perhaps for the cognate object construction (one type of Range) which may occur with die. Also, the diachronic description of starve and related hunger and thirst in Chapter 4, section 4.3.1, briefly discusses Ranged constructions that were possible for the latter two verbs. By using the term 'non-effective' the problems raised by the often used term 'inchoative' (e.g. Marantz 1984, Croft 1991, Levin 1993) can be avoided, as it is often (and originally) used in reference to inceptive verbs, e.g. Latin rubescere "begin to be red" (cf. also Haspelmath 1993:90, fh.3). In addition, the term fails to bring out the paradigmatic parallel to the transitive non-effective. I also avoid Davidse's somewhat idiosyncratic (Hallidayan) use of the term 'middle'.


45

One drawback of the term 'non-effective' is that it stands up against the deeply engrained term 'intransitive'. In response to its high entrenchment, I will mostly use this term in the descriptions. Given that within the transitive paradigm the non-effective constructions stand out much more and are much more varied than the ergative non-effective ones, it is reasonable to see the intransitive as a larger subgroup of the transitive model. Parallel to the use of 'intransitive', one could also to use the term 'inergative' for the ergative non-effective structures, and 'transitive/ergative' for the effectives (Brygida Rudzka-Ostyn, pc). While intuitively appealing, this terminology is bound to be confusing, as 'inergative' may incorrectly be identified with 'unergative' which, in the framework of Relation Grammar (cf. Perlmutter 1978) and Government and Binding (cf. Burzio 1986, Fellbaum & Kegl 1989, Levin 1993, Levin & Rappaport 1995) is used not to refer to ergative non-effectives (as I would have it) but to transitive non-effectives. In these frameworks, ergative verbs are referred to as 'unaccusatives' (hence the Unaccusative Hypothesis). For this reason, I thought it better to set up the distinctions in terms of effectiveness in which, for one-participant constructions, the effectiveness is negated. A last divergence between Davidse's terminology and mine concerns her use of 'transitive absolute' to refer to a transitive construction without over object, e.g. Soldiers trained to kill. She stretches Langacker's notion of 'absolute construal', a construal which evokes or profiles the conceptually autonomous core, i.e. the autonomous thematic relationship (cf. above).20 While in earlier work I took over this term, my attention has been drawn to its possibly confusing character, as it resembles too much the term 'absolutive' of traditional typology. Hence, I have resorted to a more common terminology, referring to the construction as an 'objectless transitive'. It should be noted, however, that 'objectless' refers to the formal aspect of the construction; as said, the construction still resorts under the effective constructions as it still has an understood object.21 By way of recapitulation, Table 2 lists some of the different instantiations of the two paradigms and aligns my terminology with Davidse's and that of Relational Grammar and Government and Binding (where relevant).

20

The term 'absolute' has been used in this sense before, e.g. Jespersen (1927) or Feltenius (1977:14). 21 At a point at which it is too late to review the entire manuscript of this book, the term 'objectless effective' was suggested to me as a more accurate cover term, making the effective character of the construction explicit. As said, any reference in this work to an 'objectless' construction should be understood as essentially effective.

46


To sum up, Davidse's work reveals how, in the domain of material processes, the relations between participants and processes are assembled differently. As such, her views are quite compatible with Langacker's approach sketched above. Yet, unlike Langacker (and many others), Davidse is not biased towards a single model of 'transitivity' but wakes us up to the many subtle distinctions in what she calls the truly 'Janus-headed' grammar of actions and events. Davidse (1991) effective

RG&GB

non-effective or intransitive

middle or intransitive

unergative

Jill died

pseudo-effective: Ranged

pseudo-effective: Ranged

(transitive)

He died a slow death. He ran a good race. We prowled the city.

My terminology Transitive effective

Ergative

transitive, causative

Examples John killed Mary.

objectless transitive transitive absolute

John kills.

middle

facility-oriented passive

Stale bread cuts easily.

effective

effective

unaccusative Jill drowned Jean. transitive, causative

non-effective

ergative middle or non-ergative

unaccusative Jean drowned.

pseudo-effective: Setting

pseudo-effective: Setting

(transitive)

|

The system burst a pipe. The window opened only with great difficulty.

middle

Table 2: Terminological overview The above description hardly does justice to the far-reaching implications of Davidse's work, yet it suffices as a starting-point for my own data-based investigation, which offers further refinements to her general description. 2.2.5 Present cognitive lexical-paradigmatic approach Even though more traditionally oriented typologists may be uncomfortable with the notion of 'ergativity' being applied to English, a formally nominative/accusative language, I subscribe to Davidse's paradigmatic views on the grammar of actions and events. The transitive and ergative systems are


47

cognitively real also for English, where they manifest themselves not in overt case markings, i.e. phenotypically, but at the cryptotypical level As Davidse says, "English realizes its ergativity as it were by a sleight of hand" (1991:147), by using the same verb in both effective and non-effective constructions. At the same time, however, through nominative case marking, the ergative effective realizes a schematic transitive image, as an instantiation of the AG-PROC-AF schema. Transitive constructions, on the other hand, realize phenotypically through case marking only one model, yet do so with considerably more specificity and force than the ergatives. In the final conclusions to this work, we will be in a position tο evaluate the double-imposure implied by the ergative model. However, my own analysis of verbs of killing has shown that her predominantly grammatical description needs to be further fine-tuned to the interaction of the grammatical patterns and the semantic structure of the verbs that instantiate them. The constructional potential of a particular verb cannot be stated as absolute (i.e. as either transitive or ergative), but depends on the kind of participants that are involved in the process. So, it is not the verb as such, in état pur, which evokes the transitive or ergative model, but the entire clause construal. Davidse does acknowledge this when she points out that clauses like in the above examples (22) John opened a tin of baked beans and (23) Mother changed the baby no longer allow the ergative effective. These sentences, Davidse (1991:150) says, realize only the transitive goal-achieving model, as the entities realizing the role of the Medium are no longer entities perceived as capable of sustaining and self-instigating the process. Although Davidse notes this constructional variation, her overall description remains pitched at the grammatical level, i.e. aimed at unravelling the "two deep 'persuasions' in the grammar of material processes" (1991:146). Slightly oversimplifying, her description gravitates too much to the top-down model, viz. it sees the general models of grammar too much as imposing their structures onto the instantiating verb. However, a unified description of the grammar of processes and events requires the inclusion of a lexical perspective, in Hallidayan terms, the more "delicate" level of grammar. Accepting the overall validity of the workings of the transitive and ergative paradigms in the grammar of causative constructions, the present work furnishes such a lexical perspective by presenting a corpus-based analysis of verbs of killing. The lexical perspective is described to the fullest in the case studies in Part II. It offers a meticulous analysis of the lexical flexibility (or the absence thereof) of verbs, which has a (direct) impact on the verb's constructional properties. Bringing the lexical perspective into the analysis can, for instance, explain why in certain usages a typically transitive verb occurs in an ergative construction and vice versa. Moreover, as our analysis of the middle

48


construction in Chapter 4 will show, verbal meaning can explain why verbs belonging to a certain paradigm (i.e. the transitive) do not occur in a construction typical of that paradigm (the middle construction). Finally, the analyses in Part II further substantiate how the ergative/transitive opposition seems to be strongly motivated by our experience of the world. More illustrations of the dynamic interplay between the two paradigms and the complex interaction of constructional and lexical meaning will certainly emerge from the corpus-based analyses, yet before broaching these we should more carefully delineate the supra-lexical category of verbs of killing as well as present our cognitive view on lexical meaning.

CHAPTER 3 THE SEMANTICS OF CAUSATIVE VERBS

Before we can proceed with the discussion of the data illustrating the lexical and constructional interplay, it is essential to have a good understanding of the semantic structure of verbs. We view the latter as forming a complex category of related senses, which encompasses a cognitive view on categories as structured around a prototype from which different meanings extend. The extension mechanisms relevant here are metaphor and metonymy which are distinct yet often interlocking mechanisms, metaphor often being mediated by metonymy. It should be noted that in this chapter only the main mechanisms will be explained (using the general verb kill) as the case studies provide further illustrations of specific lexical semantic extensions. The data on which the present study is based are drawn from one lexical domain, the field of killing. A delineation of this field is offered in the second section of this chapter, remedying some inaccuracies in Levin's (1993) classification of verbs. One of our main points is that lexical fields should in fact be regarded as supra-lexical categories which are also structured around a prototype and show fuzzy boundaries. Most people would agree that the occurrence of a verb in different constructions (e.g. the ergative effective/non-effective alternation) gives rise to related meanings. A lexical semanticist will be particularly interested in finding out whether the relatedness is a case of polysemy or vagueness. A brief commentary on this issue is offered in the last section, which also considers the issue of certain constructions being blocked in some uses of a given verb. 3.1 The structure of lexical categories Polysemy, the association of multiple, related meanings with a single form, is common with most lexical items, especially with verbs. Fellbaum points out that verbs have a higher polysemy than nouns because "verbs change their meaning depending on the kinds of noun arguments with which they co-occur, whereas the meanings of nouns tend to be more stable in the presence of different verbs" (Fellbaum 1990:278-9). Gentner & France (1988) talk about the "high mutability" of verbs. On the basis of experimental studies they show that when a verb combines with nouns that violate its selection restrictions, it is the verb's meaning that is adjusted and not that of the argument nouns. The degree of

50


adjustment depends of course on the deviation from the verb's 'normal' or prototypical use. Langacker terms the adjustment that occurs when structures combine to form composite structures 'accommodation'. He points out that the existence of composite structures and the accommodation that they entail are necessary for the simple reason that "linguistic convention cannot provide a fixed, unitary expression for every conceivable situation that a speaker might wish to describe" (1987:278). The simplest form of accommodation is when abstraction is made from specific individual situations. Consider the following examples: (1) (2)

... to condemn the killing of U.S. Lt. Col. William Higgins, who was kidnapped and killed by a Shiite terrorist group ... (WSJ) Yusef Hawkins was killed last week when he went to a largely white Brooklyn neighborhood to buy a used car. (WSJ)

The two sentences report on different events, involving different participants and circumstances; consequently, the meaning of kill is not identical (for the sake of exposition I ignore the grammatical differences). The applicability of kill to code both conceptions is the result of a process of generalization, in cognitive grammar terms: schematization, where a schema is defined as "[containing] what is shared among the more highly specified meaning, yet [...] abstract enough to be neutral with respect to the differences among them" (Lindner 1982:306). In other words, the conceptions encoded by the two sentences are specific instantiations of the schematic meaning that is the prototype of kill, viz. "to act upon a human being so that this human being dies". As is intuitively clear, kill readily extends from this prototype to encode the killing of living beings other than humans, as in the following examples: (3) (4) (5)

... the female spiders that kill [males] after mating ... (WSJ) ... bacteria that attack and kill weeds ... (WSJ) ... products that will be cooked enough to kill bacteria ... (WSJ)

These uses are no longer fully compatible with the schematic prototype of kill. In technical terms, they are only partially sanctioned by it, as the prototypical notion of 'humanness' is suspended (cf. Langacker 1987:68fi). All these uses of kill form a semantic network of related meanings, which can be represented as in Figure 10. The degree of prototypicality is represented by the degree of the thickness of the box; the full arrows with black heads represent instantiation, the dashed arrows with grey heads signal extension. It should be added that each instantiation relates to schema via extension, since specific characteristics of the instantiations are suspended to arrive at the superordinate schema. This schema covers the same conceptual territory as its instantiations but in less detail. The

THE SEMANTICS OF CAUSATIVE VERBS

51

bi-directional arrow linking the two instantiations of the prototype signals that neither of the two instantiations subcategorizes the other as they are "sisters" (Tuggy 1988) under the prototype schema.

Fig. 10: Partial semantic network for kill The meanings of kill discussed so far are all literal uses, i.e. they are characterized relative to the source domain of kill, which is represented in Figure 10 by the dashed box. Metaphors arise when the extension crosses the domain boundary. For kill, this is most overtly the case when the entity affected is not a living being, as in the following examples: (6) (7) (8)

... unfair competition that killed its cookies ... (WSJ) ... they do not want him to kill the peace process .... (WSJ) ... the plan to kill Shoreham ... (WSJ)

These usages are no longer characterized relative to the domain of kill and are thus metaphorical. It should be noted, however, that the boundary between the literal and the metaphorical usage is not always well-delineated. Like many other categories, the category of living beings has blurred edges. Regardless of possible scientific criteria defining the category, ordinary speakers will vary in their demarcation of the category. A case in point is the abortion debate between the pro-lifers and the pro-choicers. For instance, the following example may be difficult to categorize: (9)

... [the anti-conception pill RU-486] kills the unborn ...

52


Although the use of unborn implies a form of life, it is a mode of living that is different from that of the prototype. Stating it simplistically, we can say that the debate boils down to whether the unborn is considered a full member of the category of human beings or only a borderline case. The pro-lifers regard a fertilized egg as a human life form and, as a result, abortion equates killing a human being. The pro-choicers, in contrast, consider it an organism having only human life potential, and to them abortion is merely dispensing with something unwanted. In the case of abortion, the variability in categorizing the patient is often reflected in the expressions for the patient that either downplay or emphasize its 'humanness' (cf. Chapter 7). Example (9) is revealing in yet another way as it involves a metonymical shift of focus to an earlier state in man's existence. This state is situated in a different mode of living. Consider some additional examples of metonymical extensions with kill: (10) (11)

Decreasing salinity is killing plant and animal life ... (WSJ) If this goes on, we could kill the estuary in 15 years ... (WSJ)

In example (10) it is no longer the living entity that is profiled as the affected entity, but its distinguishing attribute, viz. life itself. Example (11) takes it one step further by profiling the location in which both life and living entity are situated. In metaphorical uses these metonymies also occur, as already illustrated by the examples above: in (6) the profiled Affected is the entity (cookies), in (7) a process (peace process), and in (8) the location where the power plant to be shut down is located (Shoreham). Note, however, that there is still a subtle distinction between the metaphorical usage kill a process and the literal kill life as far as the metonymical extensions are concerned, since kill life does not involve a Part/Whole relation but a metonymical shift from the carrier of an attribute to the attribute itself. One cannot felicitously profile a particular life process as such: *The burglars killed my dog's life. Against the background of the Idealized Cognitive Model of life as a process evolving over time, the metonymical extensions for both literal and metaphorical uses can be diagrammed as in Figure 11. Obviously, the diagram is a gross oversimplification and is not to be equated with the meaning of each expression. For the sake of convenience and uniformity, I have ignored the difference of metonymy between kill life and kill a process and for the former represented life as a containing process. The failure of this process to reach its natural end point is represented by the broken line arrow underneath the three states. The metaphorical understanding of states and processes as containers is also apparent in the presentation of the entity within a box (a state). The killer has been presented as a schematic participant (its non-elaborated character is


53

indicated by the cross-hatching) and the kill-process itself has been abbreviated to only a single box, whereas in fact it is a sequence of different states over time (cf. Chapter 1, Figure 1).

Fig. 11 : A simplified network of some metonymies and metaphors of kill Following the diagrammatic conventions set by Rudzka-Ostyn (1994), metonymical extensions are represented by a squiggly line. The prototypical variant (indicated by the heavy-line box) is that of a living being in a post-natal state. Profiling the (truly spatial) location is an instance of the Container-Contained opposition, which is also definable as a metonymical Part-Whole relation. A good example of such a shift within one sentence is the following: (12)

By starving the peasantentity they have starved Polandlovation (WSJ)

54


The two construals differ in their degree of the destructive impact envisaged. As a rule, when the location is affected, everything within it is affected as well, whereas the reverse need not always be true. For examples (10) and (11), both taken from the same article, the metonymies are moreover textually significant. This is indicated by the following schema of the information flow of the immediate context (S n refers to the sentence number, cln to that of the clause): S1:

the estuary is turning into a biological desert

S 2 : cl1 : decreasing salinity is killing plant and animal life cl 2 : the goo has done in...fish species and invertebrates S3:

we could kill the estuary

location life examples location

The metonymical variation thus functions to ensure textual cohesion and argumentative force. Chapter 7 will show how similar variations in the coding of the Affected occur with abort depending on the speakers' ideological stance. As already mentioned, the rhetoric of a pro-choicer tends to feature expressions that downplay the "life" feature (such as fetus, (pre-)embryo, or pregnancy), while a pro-lifer will upgrade the "life" aspect and favour expressions like child, baby, or human being (cf. also Coulson 1992). Figure 11 shows that metonymy concerns an association based on contiguity, or as Deane puts it, "with metonymy spreading activation renders a related concept salient, qualifying it for reference" (1988:331). 22 Lakoff points out that metonymy "occurs within a single conceptual domain [...] structured by an ICM" (1987:288). Metaphors, in contrast, establish correspondences between different domains; a metaphorical extension occurs if "the path from instantiation to schema crosses domain boundaries" (Rudzka-Ostyn 1994:417). However, Figure 11 already shows that metaphor and metonymy are often interrelated. Consider the usage kill a process, which involves a combination of metaphor and metonymy, a "metaphtonymy" (Goossens 1990). The usage metaphorically maps onto the domain of processes and within this metaphorical usage, the Part/Whole metonymy leads to profiling the process, not the entity undergoing the process. The complexity of the issue can be revealed if a usage such as to abort a pregnancy is considered. It involves a metonymical extension from fetus to pregnancy, and stays within the domain boundaries. However, 22

See among others Apresjan (1973), Croft (1993), Lakoff & Johnson (1980), or Nunberg (1978; 1979) on metonymy.


55

many people consider it a metaphorical extension, as a specific instantiation of to abort a process. Among other things, the problem centres around the difficulty of delineating conceptual domains. We will not go into this issue, as this would lead us too far afield (see e.g. Feyaerts 1997). Thus far, the description has focused on variations in the entity that functions as the Affected. The 'stretching' of the selection restrictions for the Affected will typically give rise to substantial accommodation of the verb's semantics and often lead to metaphor. As said, even when a usage falls within the boundaries of the source domain but shifts the profile to an unprototypical participant, as in to kill life or to kill the estuary, the usage borders on being metaphorical. Although I have illustrated the variation of the Affected with transitive kill only, it is clear that similar metonymical and metaphorical extensions occur with virtually all verbs of killing. Significantly, variations with the Agent generally do not lead to a metaphorization; usages like be killed by stray bullets, be killed in an accident, or be choked by a coat of ash are still characterized relative to the source domain. Clearly, the reason is that when it comes to metaphorical usage, verbs of killing have a fundamental hook-up with the Affected, as their source domain matrix imposes the relatively strict requirement that this entity be a living being. But this is surely not the only dimension along which metaphorization may occur. Consider hyperbolical uses such as my feet are killing me or I'm drowning in paper work, in which the event does not refer to a 'real' death event. Variations in the type of Agent are clearly relevant to the semantics of the entire construal (see Chapter 5, section 5.1 for a more elaborate discussion), but for verbs of killing it appears to be a dimension less likely to trigger 'true' metaphorization. 3.2 The structure of lexical fields In her classification of verbs, Levin (1993) distinguishes two categories of verbs of killing, the MURDER verbs and the POISON verbs. She lists the following class members for each category: MURDER verbs: assassinate, butcher, dispatch, eliminate, execute, immolate, kill, liquidate, massacre, murder, slaughter, slay POISON verbs: asphyxiate, crucify, drown, electrocute, garrotte, hang, knife, shoot, stab, strangle, suffocate Her group of MURDER verbs comprises verbs which all describe killing, but "none provides any information about how the killing came about" (1993:231), a property which sets them apart from the POISON verbs which "lexicalize a means", i.e. 'Verbs which relate to actions which can be ways of killing"

56


(1993:232). While Levin's classification is correct for the MURDER verbs, her semantic grid is not refined enough when it comes to her class of POISON verbs, which is a relatively heterogeneous group both lexically and grammatically. Levin herself is well aware of the grammatical heterogeneity, pointing out that while the POISON verbs as a rule do not allow the ergative non-effective, some items of the group, viz. suffocate, asphyxiate, and drown do. Compare: (13)a. b. (14)a. b.

The witch poisoned Snow White. (Levin 1993) *Snow White poisoned. (Levin 1993) Somebody drowned Esther Williams. Esther Williams drowned.

Levin tries to salvage her classification by cross-categorizing the three verbs under the class of SUFFOCATE verbs (a subgroup of processes involving the body), which allow the alternation. Levin defines the SUFFOCATE verbs as "[relating] to the disruption of breathing" (1993:224) and lists the following members: SUFFOCATE verbs: asphyxiate, choke, drown, ?stifle, suffocate Although the cross-classification solves part of the problem, it is relatively ad hoc, since smother, strangle or throttle are not listed under the SUFFOCATE verbs and throttle is absent from both groups. On the basis of their semantic structure, these three verbs should nevertheless be included in the class of SUFFOCATE verbs, even though they do not share the grammatical properties of the others. She notes that the SUFFOCATE verbs "differ from each other with respect to the means" (1993:224). While this is true to some extent, Chapter 6 will demonstrate that a more nuanced statement is in order. Finally, Levin does not explain the question mark accompanying stifle. The inadequacy of Levin's characterization of the POISON verbs is due to first, an incorrect identification of the verbs' meanings (which is surprising given her commitment to verbal semantics) and secondly, her neglect of the distinction between the transitive and ergative paradigms. I propose an alternative classification of the field which does more justice to the verbs' semantics and to their place in the two paradigms. My alternative classification pertains especially to Levin's POISON and SUFFOCATE categories; I basically agree with her classification of the MURDER verbs and merely mention that lynch should be added to them. The most prominent group in Levin's POISON category is that of the SUFFOCATE verbs listed above. As said, I will include smother, strangle and throttle despite their different grammatical properties. However, there has been, and to a certain extent continues to be, some amount of reinterpretation which


57

blurs the grammatical and lexical differences. For example, the next chapter (section 4.3.2) discusses the grammatical convergence into an overall ergative pattern which occurred with the SUFFOCATE verbs in an earlier phase of English of which certain traces in present-day English can be found. The verbs' lexical convergence occurring in the same time period 'and partially continuing up to the present is presented in Chapter 6. A sibling subgroup of processes involving the body (absent from Levin's work) is that of the STARVE verbs starve and famish. As will also be shown in the diachronic study in the next chapter, from an etymological point of view, hunger and thirst should also be included, but in modern English they are to be regarded as intransitive verbs of desire (cf. Levin 1993:194) aligning with verbs such as long, yearn or desire. A third group emerging from Levin's POISON category contains the 'true' POISON verbs, for which I will use the more mnemonic label INSTRUMENT verbs. These include verbs like knife, garrotte, poison, or crucify, which are verbalizations (usually via zero-derivation) of nouns which designate instruments that can be used to kill someone. The verbalization thus adds conceptual content and the meaning of the verb is paraphrasable as "use [INSTRUMENT] to harm or kill" (cf. also Clark & Clark 1979). The INSTRUMENT verbs are further unified by not allowing the ergative non-effective. In addition to Levin's example (13) I can mention: (15)a. Otis L. Jones...was convicted of garroting his wife in 1981. (WSJ) b. * Jones's wife garroted in 1981. The INSTRUMENT verbs are, at least in principle, an open-ended set. In the proper context, any instrument used in a killing can be expressed verbally. Some time ago, I heard an arrested suspect expecting to be sentenced to death for murder say on television "I am sure going to be chaired". Next, verbs such as hang, stab or shoot designate a specific action that may lead to the death of the Affected. Consequently, they are verbs of killing by extension but retain the profile on the specific action, and hence they are called 23 ACTION verbs. They too form a separate class of verbs expressing the agent's action that has physically affected the victim, possibly resulting in the victim's death. Often, one finds contrastive pairs of action and instrument, such as to hang vs. to rope/string or to shoot vs. to (machine-)gun or to stab vs. to knife. Usually, however, the ACTION verb is more general than the INSTRUMENT verb. Thus, as far as the conceptual content is concerned, an INSTRUMENT verb 23

It will be readily understood that the label 'action verb' is only a shorthand and is certainly not meant to imply that the other 'verbs of killing' do not designate an action.

58

LEXICAL PERSPECTIVES ON TRANSITIVITY AND ERGATI VITY

evokes a more specific conception of the event than an ACTION verb. Possibly, hang could be taken as a verb straddling the border between the ACTION verbs and the SUFFOCATE verbs, since suffocation is implied. However, I will still classify it as an ACTION verb, as the profile is on the action and not on the implied subevent of suffocation. Finally, there is a fifth, admittedly quite restricted, set of verbs that Levin overlooks, namely, those that entail an additional focus on the body part afflicted, like decapitate or behead. In fact, apart from the rare verb decollate, these are the only verbs that embody a profile on the severing of the head from the body. Typically, more elaborate constructions are used to profile certain types of bodily harm, e.g. to slit one's throat or to blow someone's brains out. In summary, on the basis of lexical content I have regrouped Levin's verbs of killing into the following subcategories: 1. 2. 3. 4. 5. 6.

MURDER verbs SUFFOCATE verbs STARVE verbs INSTRUMENT verbs ACTION verbs DECAPITATE verbs

This reclassification yields a more coherent classification of the field of killing both lexically and grammatically. Figure 12 visualizes the general classification.

Figure 12: General classification of verbs of killing. The labels of the categories have been underlined; members are in smaller lettersize and in italics. While the figure is a gross oversimplification, it nonetheless reveals a great deal. Most obvious is the transitive predilection of


59

verbs of killing. The reason is straightforward: the concept of killing someone is quite compatible with the meaning of the transitive paradigm, in which an inert Goal is affected by the action of a volitional Actor. Chapter 5 will demonstrate that the MURDER verbs are fundamentally related to the Actor's intentionality, and are thus prototypically transitive. Moreover, they are also prototypical within the field of killing, as indicated by the thickness of the circle, and their salience radiates to the whole field. The prototypically transitive character of the INSTRUMENT verbs also finds a plausible explanation against the background of the Actor's intentionality (see Davidse 1991:60-1). Following DeLancey (1984b), an instrument has to be viewed as an extension of the Agent's will and thus presupposes intentionality. Verbs derived from instruments are thus bound to be transitive, yet they are less prototypical verbs of killing. The transitive character of the DECAPITATE verbs is tied to the type of bodily harm: a severage of head from body is a process not likely to be conceived of as coming about spontaneously, but strongly evokes the notion of the agent inflicting the harm on the victim (cf. also Haspelmath 1993:94). The INSTRUMENT and DECAPITATE verbs further enforce the transitivity of the field. The STARVE verbs are typically ergative as are the SUFFOCATE verbs, although the latter comprise three transitive members {strangle, throttle, and smother). The paradigmatically complex evolution of starve (and related hunger and thirst) is discussed in Chapter 4, section 4.3.1. Chapter 6 explains the ergative predilection of the SUFFOCATE verbs as experientially motivated, i.e. grounded in our experience of the world. It will further reveal how the distinction between the transitive and ergative SUFFOCATE verbs has an experiential basis as well. There are two important nuances to be added to the above description of the field of killing. The first concerns the structure of the field. In Figure 12, the field is presented as nicely carved up into distinct, well-delineated categories. While this presentation is acceptable for explanatory purposes, it is not entirely accurate. The non-equality of the distinct subgroups has already been observed: the MURDER verbs are better representatives of kill-expressions than the SUFFOCATE or the STARVE verbs. As already mentioned, one reason for this is that the former accord well with the general conception of killing as a process in which a passive victim is affected by the act of an intentional Agent. Another reason is the salience of goal-achievement in the semantics of the MURDER verbs. In contrast, for the SUFFOCATE or STARVE verbs goal-achievement is much less prominent (see Chapters 4 and 6). Within each subgroup, there are also prototype effects. Stifle and asphyxiate are less typical members of the SUFFOCATE verbs, and chair, as used in I am sure going to be chaired, is a peripheral member of the set of INSTRUMENT verbs. Apart from the internal differences in salience within the field and within each subgroup, the boundaries of the subgroups cannot always be sharply

60


delineated. For instance, hang has been classified as an ACTION verb, yet as said, it shares certain components with the SUFFOCATE verbs, viz. the constriction of the throat and the consequent lack of oxygen. Thus, it is not surprising to see that in an older phase of English, when doublets were still common, hang was often paired with other SUFFOCATE verbs, e.g. hanged and strangled, hanged and stifled (see Chapter 6, section 6.2, on such doublets). In brief, the lexical field of killing emerges as a prototype-based category whose external boundaries are not sharply delineated, and whose internal structure shows considerable flexibility and differences in salience (see also Geeraerts et al. 1994 for a systematic exploration of these properties of lexical fields). The second nuancing to the above description concerns the already noted variability that may occur with one and the same verb. Presented as it is now, the lexical and constructional classification of verbs assumes that the paradigmatic address of a verb is fixed and stable. In Chapter 2, section 2.2.3, Levin (1993) was criticised precisely for adopting such an absolute view on verbal semantics and ignoring the verbs' polysemy which may affect the type of alternations allowed. I thus also disagree with Talmy, who claims that the diathesis alternations (which he terms 'lexicalization patterns') are" "associated with particular lexical items" (1985b:84). Talmy's insights are inspiring, yet like Levin he neglects that an item's lexicalization patterns are not fixed and stable and may vary contextually. As with any cognitive category, prototypicality is involved here: as said, some verbs are more prototypical instantiations of a paradigm than others. However, given the dynamic nature of the clause construal, deviations from the prototype are to be expected. In fact, it is incorrect to talk about ergative or transitive 'verbs', whether in the present framework or in another. Rather, what is at issue is their transitive or ergative use (cf. also Jespersen 1927 or Feltenius 1977). However, given that verbs are prototypically associated with one paradigm, it is partially correct to talk about transitive or ergative verbs. In reference to this paradigmatic prototype I will, for the sake of convenience, often use this shorthand. In short, both lexical categories and lexical fields emerge as prototypestructured categories showing degrees of salience and fuzzy boundaries which leads to grey transition areas. To this, another layer of prototypicality can be added, viz. that of the transitive and ergative paradigms themselves; they also represent conceptual categories structured around a salient structure. Their prototype structure has been pointed out by Davidse (1991), drawing on insights from Cognitive Grammar. Within each paradigm certain structures are more prototypical than others. Logically, then, and in line with the observations on verbal semantics above, alternation patterns cannot be seen as absolute


61

criteria, applying to all instances of a category. The following diagram (taken from Davidse 1991:156) represents the prototypical structure of the two paradigms:

Examples

; Legend I

instantiation extension

I

prototype

|

mapping

pseudo-effective

(a) It is raining.

(h) The bomb killed Mary.

(b) John fell.

(i) The floor paints easily.

(c) John is running.

(j) Ethiopian children starved.

(d) The river flows.

(k) The general marched the soldiers,

(e) The car moved.

(I) The government starved children.

(f) John accidentally hit Mary.

(m) John grew a wart.

{ (g) John killed Mary.

(n) John sang a ballad.

Fig. 13: Simplified schematic network of transitive and ergative categories Davidse sees as the prototypical instantiation of the transitive effective the volitional goal-directed action (g) and that of the ergative effective, the instigation of process (k). For the non-effective constructions, the controlled intransitive process with a volitional Actor (c) is the prototype within the

62


transitive paradigm, and the ergative non-effective (j) is that for the ergative. In the non-effective schema, the 'ActorMedium' is the general participant that schematizes over the intransitive Actor and ergative Medium; it lacks the features of true agency or real affectedness.24 Clearly, the diagram shows more than has been discussed so far and includes patterns which are not directly relevant to the present study. Anticipating the discussion in the next chapter, it can be pointed out that the ergativization of intransitives (e) as well as the possible reinterpretation of a transitive middle construction (i) as an ergative non-effective (j) have been represented as extensions crossing paradigm boundaries. In Figure 13 there is considerable overlapping of the transitive and ergative paradigm, just as in Figure 12. However, the overlapping is different for each: in Figure 12 it serves to capture the cross-paradigmatic nature of some subfields of the field of killing, whereas in Figure 13, it concerns the internal structure of the paradigms themselves. Against the background of the preceding discussion, the considerable complexity involved in the study of lexical and constructional meaning becomes clear. Suffice it to give a simple example to illustrate the three layers in our descriptive analyses. The MURDER verbs can be regarded as prototypical within the field of killing, they are prototypically transitive, and thus they show up in constructions that are prototypical within the transitive paradigm. It can be noted that our analysis of the MURDER verbs (Chapter 5) suggests a refinement as to the Davidse's characterization of the transitive prototype. It will also illustrate that kill, the verb most prominently present in the literature as a representative of 'transitive' verbs, is in fact the least transitive MURDER verb, as it is less tied to the transitive prototype. 3.3 Lexical versus constructional semantics Chapter 2 argued that one cannot begin to accurately describe the semantics of verbs without taking into account the semantics of the constructions in which they can occur. The constructional variation that is possible with verbs adds an important dimension to their semantic flexibility. On the other hand, the descriptions in this work reveal that a proper characterization of constructional semantics necessitates an accurate analysis of verbal meaning. In view of the network conception of verbal semantics, how is constructional variation to be interpreted? Does constructional variation give rise to clearly distinguished 24

Davidse uses the (quoted) term 'Medium' to refer to this generalized participant which is an unfortunate choice inviting misunderstanding. Hence, I have opted for the compound term 'ActorMedium' (abbreviated as ACME). It can be noted that this schema is highly schematic and thus less salient, certainly in comparison to the effective AG-PROC-AF schema.


63

meanings (polysemy) or does it merely represent different usages of a single general meaning (vagueness)? Or should the question be reversed: is it a difference in verbal meaning which triggers constructional variation? While at this point it is still too early to come to clear answers, some exploratory thoughts on these issues can be offered. The 'modulation' of a verb's meaning due to its occurrence in different constructions is intuitively characterized as leading to semantically related uses. For example, the two usages of the verb suffocate in the constructionally related sentences John suffocated Mary and Mary suffocated are usually perceived as semantically related. Consequently, the relationship between the two variants can be identified as 'constructionally determined polysemy'. From a lexical semantic point of view, the question arises whether the semantic multiplicity is not so much a case of polysemy but rather an instance of vagueness, in which case the different usages of the verb are to be regarded as non-distinguished subcases of a single, more general meaning. In traditional terms (e.g. Lyons 1977, Zwicky and Sadock 1975), this would involve only one lexeme with a single but non-specific meaning, whereas in the case of polysemy one lexeme with two distinct senses would be involved. Not that in the case of homonymy, there would be two distinct lexemes, as there is no semantic relationship between the different meanings. However, various studies have shown that the lexical-semantic distinction between polysemy and homonymy, or ambiguity and vagueness is problematic.25 The reason is that the criterion used to distinguish between these notions, viz. the relatedness of meaning, is itself variable. In answer to this vexed problem, Tuggy (1993) has furnished a coherent cognitive account by postulating a continuum with homonymy at one end (maximal meaning distinctness) and vagueness at the other (maximal meaning relatedness). Polysemy straddles the border, as it involves meanings that are distinguishable, yet also related (cf. also Deane 1988). Applied to the kill-examples above, one can say that the meanings in (1) (the killing of U.S. Lt. Col. William Higgins) and (2) (the killing of Yusef Hawkins) are instances of vagueness rather than polysemy. Tuggy explains vagueness as involving "meanings which are not well-entrenched but which have a relatively well-entrenched, elaborately close schema subsuming them" (1993:280-1). This is clearly represented in Figure 10 above, in which the low-level schema "act upon a human being etc." is indicated as having a higher salience than the instantiations. The structure of the network also reveals that the greater the distance from the prototype, the less related the meanings will be. In those 25

E.g. Cowie 1982; Cruse 1982, 1986; Deane 1988; Geeraerts 1993; Lehrer 1974; Lipka 1986; Nunberg 1978, 1979; Tuggy 1993.

64


cases, the schema linking the instantiations will be much more schematic and less entrenched than the individual meanings. Metaphorical extensions will thus be experienced as more polysemous and will even lead to ambiguity sometimes, as indicated by the following zeugma when two readings of kill, viz. "act upon a human being etc." and "pass time" are conjoined: (16)

John kills and so does Mary.

Although related, the two meanings of kill are so remote that one can, in normal usage (i.e. disregarding puns), no longer felicitously combine these readings without semantic oddness. If the issue is already complex for 'lexically determined' semantic multiplicity, it becomes even more complex when 'constructionally determined' semantic multiplicity is brought in, as it adds another dimension of variation.26 Also here, the degree of relatedness between different constructions may vary. In the typical case, constructional variation is situated closer to the vagueness end, especially with the so-called 'regular' alternation patterns. For instance, in the context of the effective/non-effective alternation common with ergative verbs, the two readings of starve in Poland is starving and the Communists are Poland are closely related, differing only in the resolution of the instigation by the second clause. From a (narrow) lexical semantic point of view, the relation between the two could be seen as an instance of a Part-Whole metonymy, since the non-effective profiles only part of the larger conceptual base. Similarly, an objectless transitive, e.g. Soldiers trained to kill, could be said to single out only a part of the transitive base process (see Chapter 5, section 5.4). Such a view on alternations is only partially correct (i.e. within an exclusively lexical perspective), since it falls short of observing the paradigmatically determined systematicity of such 'metonymies'. Further, it disregards the underlying semantics of the constructional variation, which in both examples is quite different. For less typical constructional variation, the degree of relatedness between the different structures is usually less strong. In these cases, it is more appropriate to regard the different constructions as polysemous or even homonymous. For example, the next chapter discusses some unusual ergativizations such as the engine throttled down to idle. Chapter 7 reveals the constructional split with present-day abort as determined by its usage. In these cases, certain constructional alternations are restricted to one particular type of usage. The constructional extension from the prototype in fact reinforces the Most studies, typically focusing on semantic multiplicity with nouns or adjectives, ignore the constructional dimension altogether.


65

lexical semantic extension. In other words, in both cases the metaphorical extensions are further differentiated from the source domain by a systematic paradigmatic opposition. So, in these cases, the extension gives rise to more full-fledged polysemy rather than vagueness. In the traditional view, the syntagmatic differences can be seen as a clear test for polysemy: one meaning of a verb can occur in a syntagm in which another cannot. Note that while most speakers will still regard abort as a polysemic cluster (despite the paradigmatic opposition between literal and metaphorical uses), it appears that for throttle, the lexical and constructional evolution has caused many speakers to perceive the two uses as homonymous. From the above description the view emerges that the issue of semantic relatedness is not a simple matter, for the lexical as well as for the constructional flexibility of verbs. So far, however, the discussion has treated the two dimensions of variation as separate phenomena, whereas in fact, they interact in intricate ways. In the beginning of this chapter, it was observed that verbs have a higher degree of polysemy than nouns, because their meaning depends on the kinds of arguments they take. Importantly, the semantic 'accommodation' triggered by the type of arguments does not only affect the 'lexical' meaning of a verb, but also its constructional possibilities. For example, it will be recalled from Chapter 2 that Davidse (1991) mentions that ergative verbs, like open and change, may surface in constructions that no longer instantiate the ergative model, as they no longer allow the non-effective alternation (Chapter 2, examples (22) John opened a tin of baked beans and (23) Mother changed the baby). In such clauses, the notion of co-participation is no longer present: the Medium is not conceived as capable of sustaining and self-instigating the process. Hence, they realize only the transitive model. Davidse's views are correct, yet the present study takes the issue one step further, drawing specific attention to the lexical semantic flexibility of verbs which may trigger paradigmatic changes. In other words, I take a view on clausal semantics which goes much further than Davidse's, not restricting it to the 'accommodation' of the verb by the argument nouns. The descriptive parts of this work reveal how lexical and constructional semantics are interdependent in intricate ways. Clearly, when one says that a prototypically ergative verb may surface in an exclusively transitive construction, the question arises as to whether ergativity has any cognitive value at all, especially when it comes to effective constructions. In other words, would it not be easier to say that even effective clauses that allow the non-effective alternation (e.g. John suffocated Mary vs. Mary suffocated) are not ergative but transitive. Ergativity would then be restricted to the non-effective clauses with these verbs. This position has been common in the linguistic literature.

66


However, despite the impossibility of the ergative non-effective alternation for some uses of prototypically ergative verbs, I subscribe to the idea that a true ergative effective semantically differs from a transitive effective, even though this difference may not always be immediately visible in the surface structure. It is too early at this point to critically evaluate this complex issue at length. What is crucially involved is the intricate semantics of the clause construal in its entirety, which involves the dynamic interaction between the semantics of the construction and that of the instantiating verb. The analyses in Part II of this book will provide ample illustration of the constructional variability as determined by specific uses of the verb. After these descriptions, I will be in a position to consider from a more theoretical point of view this dynamic interdependence as well as the relationship between the transitive and ergative paradigms. In these final conclusions, I will also consider other recent linguistic research (e.g. Levin & Rappaport 1995) which shows some agreement (yet also some divergence) with ours. 3.4 Conclusions to Part I More and more, the linguistic scene is opening up to realizing the complexity of clausal semantics. Former purely syntax-based models are waking up to the complexities of lexical semantics, a move which may have been enhanced by the emphasis that Cognitive Grammar has put on the inherent meaningfulness of linguistic structure. Conversely, former purely lexically oriented models have come to realize the importance of grammatical patterns as meaningful units in and by themselves. The present work situates itself in the middle of this (careful) rapprochement, arguing that it is essential to consider the semantics of the entire clause in which a verb is used. My view is that one has to refrain from adopting either a purely lexical approach that focuses unilaterally on verbs in état pur or a purely grammatical point of view which ignores the impact of verbal meaning on a verb's constructional potential. The analyses in the next part present in more detail my investigation into the field of killing and try to shed light on the complex interaction between the two models of causality, the transitive and the ergative, and verbal semantics The four chapters of Part II each have their own profile, in accordance with the issues central to them. To recapitulate, the issues discussed are the following. Chapter 4 provides a synchronic and diachronic perspective on how the lexical structure of verbs may either constrain or encourage constructional extensions. The synchronic description deals with constraints on the middle alternation as manifested by verbs of killing. It also draws semantic parallels to morphological derivations following the same pattern. The second section in the chapter offers a diachronic perspective on a number of paradigmatic shifts. Some of these have been lost again in the evolution to present-day English (e.g.


67

the ergativization of hunger), others have left some isolated traces (e.g. smother) and still others have entrenched as a new constructional prototype (e.g. starve). Chapter 5 is concerned with a more accurate description of the transitive paradigm. The different perspectives in the chapter, predominantly (but not exclusively) based on the MURDER verbs, will substantiate that the conceptual centre of the transitive paradigm is the Agent, whose action is targeted at an inert and fully affected Goal. Logically, the chapter opens by clarifying the experientially motivated prototype structure of the Agent-category. The rest of the chapter further illustrates the Agent-centredness of the transitive paradigm in general and the MURDER verbs in particular, by looking at it from a lexical, morphological and constructional perspective. The lexical perspective explores how the Agent-centredness is in fact 'hardwired' into the semantic structure of the MURDER verbs. The morphological perspective (section 5.3) offers new insights into the agentive noun formation (via the -er suffix) by demonstrating that the meaning and the distribution of the -er suffix are governed by the transitive and ergative paradigms, with the former being the more dominant one. The constructional perspective in section 5.4 is concerned with the objectless transitive and argues that a proper characterization of its usage can be given only by taking into account the semantics of the construction itself, i.e. its maximal focus on the Action-component. Chapter 6 furnishes the ergative counterpart to Chapter 5, as it discusses the predominantly ergative SUFFOCATE verbs. The ergative paradigm gravitates towards the Affected (the Medium), which is conceived as capable of sustaining and self-instigating the process. The first section of the chapter reveals the experiential basis of the ergative predilection of the SUFFOCATE verbs, pointing out that the focus on the co-participation of the Medium is strongly conditioned by our experience. The second section discusses the enormous lexical flexibility and overlapping that characterizes the evolution of the SUFFOCATE verbs. The section thus provides the lexical counterpart to the earlier description of the constructional convergence, viz. the ergativization that occurred in the same period. In present-day English, however, the SUFFOCATE verbs show less lexical and construction overlapping, and this process of differentiation is treated in section 6.3. The last section of the chapter considers a particular phenomenon of clausal semantics, viz. how particles may influence the constructional potential of the ergative SUFFOCATE verbs drown and choke. Chapter 7 focuses on the interplay between the transitive and the ergative paradigm in the polysemic cluster abort. The transitive/ergative split, it will be argued, is experientially motivated, in which not only the human interaction with the world is relevant but also the ideology living within the speech community.

68


The complexity of the issue is illustrated by tracing the item's lexical and grammatical evolution from its introduction into English up to the present-day. It can be noted that the descriptions in Part II do not aspire to offer a fullfledged field analysis but is set up as a investigation into the issue of lexical and constructional interaction. The restriction to one field has been a practical choice, since it would be impossible within the scope of this book to present a detailed study of the interaction of grammatical and lexical meaning for the entire lexicon. Indisputably, the present research should be complemented with simñar investigations for other lexical fields in which different factors may condition constructional flexibility.

PART II

CASE STUDIES

CHAPTER 4 LEXICALLY DETERMINED FLEXIBILITY AND CONSTRAINTS

4.1 Introduction The present chapter presents two descriptions which, from a lexical point of view, are relatively general descriptions as they concern various verbs within the field or even the entire field. As such, the chapter ensures a more smooth transition from the more general Part I to the following chapters in this part which have more specific lexical ties to certain subsets within the field. This chapter offers both a synchronic and a diachronic perspective on lexical and constructional variability. Both descriptions reveal lexical meaning is crucial to an accurate account of a verb's constructional possibilities, yet they do not lose sight of the semantics of the constructional variation itself. As such, the analyses present a first exploration into the complex interaction of lexical and constructional meaning. Both descriptions further have in common that they are less tied to the contemporary text corpora. Obviously, for the diachronic description this is by definition the case; it is based primarily on historical data from the OED. More interestingly is that the synchronic description, dealing with lexically determined constraints on the middle construction, is negatively related to the contemporary corpora. That is, the innovative insights on the middle construction offered in this chapter have been inspired by the absence of the construction in the extracted corpus of verbs of killing (we will thus define them as prototypically 'non-middable'). As such, the analysis provides a nice illustration that even when corpus material does not positively confirm the occurrence of a pattern, it can shed light on the semantics of this pattern itself. This is all the more true for the middle construction which one initially would have expected to occur with verbs that are prototypically transitive. 4.2 A synchronic perspective on constructional constraints Much has already been said about the middle or medio-passive construction, yet very often in conflicting terms. Section 4.2.1 tries to offer a proper characterization of this construction. It opens by considering some relevant observations that have been made by researchers from various theoretical

72


backgrounds.27 This discussion is somewhat selective, as different researchers have made similar comments. Also, some issues are discussed at more appropriate places later in the description. In a second move, we consider Langacker's cognitive analysis of the middle construction which, as already mentioned in Chapter 2, section 2.2.2, needs to be refined by a paradigmatic account as offered in Davidse (1991). In turn, Davidse's account can be refined in two ways, as elaborated in section 4.2.2. First, the slight transitive bias of her description should be remedied to come to a more careful statement on the transitive/ergative interplay regarding the middle construction (section 4.2.2.1). Secondly, the lexical meaning of a verb should be considered as well, since it determines the acceptability of the middle construction (section 4.2.2.2). Our lexical-paradigmatic account is further confirmed by semantically parallel restrictions for morphological processes, such as the -able derivation (section 4.2.2.3). 4.2.1 Defining the middle construction Keyser & Roeper (1984) present a host of evidence which signals the difference between middles and ergatives. They observe that middle constructions "state propositions that are held to be generally true [...] they do not describe particular events in time" (1984:384). Similarly, Fagan notes that "middles [...] are not used to report events, but to attribute a specific property to some object" (1988:200). Consequently, Keyser & Roeper say, middle constructions, e.g. Greek translates easily, do not allow the imperative (*Translate, Greek!), whereas 'true' ergative verbs do, e.g. Sink, boat! or Bounce, ball! Further, the middle construction necessarily implies an agent: they "state the doability of an action as it can be, or is, performed by a non-specific agent, i.e., anybody" (Fellbaum 1985:29). The notion of feasibility and that of the implied agent have been observed in the literature, e.g. Fiengo (1980) notes that middles retain an implicit agent whereas ergatives do not, Levin (1993:5) talks about "an understood but unexpressed agent". It is a point also made by Langacker, who explains the middle construction against the background of the action chain, saying that it profiles the conceptually autonomous core, the "thematic participant" (see Chapter 2, section 2.2.2). He says about the following sentences (1) (2) 27

The window opened only with great difficulty. (Langacker 1991b) The dried mud scraped off effortlessly. (Langacker 1991b)

See e.g. Fagan 1988; Fellbaum 1985, 1989; Fellbaum & Zribi-Hertz 1989; Fiengo 1980; Hoekstra & Roberts 1993; Jespersen 1927 (calling it the 'activo-passive'); Kemmer 1993;

LEXICALLY DETERMINED FLEXIBILITY AND CONSTRAINTS

73

that "while the ease or difficulty of carrying out the action is attributed to inherent properties of the subject, it can only be assessed as easy or hard in relation to the capacity of an actual or potential agent" (1991b:334). The conceptualization designated by these sentences indeed incorporates the agent as an unprofiled part of the scope of predication. However, as should be clear by now, open need not occur in a middle construction to single out the thematic participant. The non-effective the window opened also profiles this participant, which is not possible with scrape off: *the dried mud scraped off. Further, while (1) may still strongly evoke an agent, this need not always be the case with an ergative verb, even when a "facility adverb" (Vendler 1984) is added to it (cf. below). Although Langacker (1991b) acknowledges that the middle construction implies an agent, he nevertheless subscribes to Van Oosten's (1977) view that the subject is portrayed "as being in some sense responsible for the profiled action" (Langacker 1991b:334, my emph.). This view is shared by a number of people: Ryder talks about an "agentive patient" (1991b:309); Kemmer talks about "the Initiator status of the Patient" which "derives from the fact that the event is conceived of as proceeding from the Patient" (1993:147, my emph.). Although it is true that the middle construction shifts the focus to the Affected participant, the voice of the construal remains passive, as illustrated by Smith's (1978) do-test, illustrated in example (3): (3)a. The story reads quickly. (Smith 1978:102) b. *Did the story read? The ungrammaticality of (3b) signals that the subject of the construction does not become an agentive participant: it does not do anything, but something is done to it, and it is said that this 'doing to' is easy. Davidse (1991) defends a similar position, continuing Halliday's views that middle clauses are GOAL-PROCESS constellations in which "[the Goals] do not actively co-participate in the process as Mediums do [...], they are affected as inert participants only by the Actor's action" (1991:42). In other words, while the middle construction shifts the focus to the Goal, it does not change the 'inherent voice', which remains passive. In this view, Davidse's use of Fawcett's (1980) term "facility-oriented passive" makes good sense. Comparably, Faltz (1977) uses the term 'facultative'. In her description of the middle construction, Davidse is primarily concerned with the paradigmatic opposition between transitive middles and ergative nonKeyser & Roeper 1984; Lakoff 1977; Van Oosten 1977 (introducing the term 'Patient-Subject construction') and Williams 1981.

74


effectives. She further points out that Lakoff's (1970:250) example Sheila seduces easily and willingly is ambiguous (an ambiguity that has escaped Lakoff). Either this sentence is interpreted as a middle construction saying that Sheila is the target of someone else's act of seducing, and it is said that it is easy for this implied agent to be successful, or it is interpreted to mean that she is the agent of the seducing, and that she does this easily and willingly. Note that when the order of adverbs is reversed, Sheila seduces willingly and easily, only the latter reading with Sheila as volitional agent is possible. Fellbaum (1985:26, fn. 1), arguing against van Oosten (1977), makes exactly the same point. 4.2.2 A lexical-paradigmatic view on the middle construction 4.2.2.1 Transitive versus ergative middle constructions. While I believe that Davidse's views are essentially correct, I have some reservations concerning her commitment to "narrow it down to a strictly transitive phenomenon" (1991:42), since I see an ergative side to the story as well. In fact, ergative constructions show an ambiguity that complements the one she has observed for transitives. More specifically, the adverbial modification in a sentence like the door opens easily may comment upon the facility of either someone opening the door (modification of the '(IMPLIED ERG)-open' unit) or the door opening itself (the 'ABS-open' unit). In other words, the voice vagueness of the non-effective is retained here. However, some verbs that display the ergative patterning tend to foreground the modification to the 'ABS-process' when a facility adverb is added. Consider (4) (5)

Frogs drown easily in improperly oxygenated water. Asthmatic patients choke easily.

In these sentences, the modification of the instigation, while not fully excluded, is secondary. The paradigmatically determined ambiguities can be represented as in Figure 14, where E stands for 'entity' and FC, for 'feasibility comment'. Following Langacker (1991b), the feasibility is represented by the symbol >. In the presentation of the ergative constructions, the representation of the feasibility has been turned 90° (Λ), iconic with the nuclear orientation of the ergative paradigm. While facilitation is most common, hindrance (< or v) may also be at issue in the middle construction, as in (c), which may culminate in the negation of the doability of the process as in This dress won't fasten (Fagan 1988:201). Note that in Figure 14 only (a) and (c) represent transitive and ergative middle constructions; (b) represents an objectless transitive construc tion with a facility adverb added to it; and (d) is an ergative non-effective (with minimal implication of an Instigator), again with a facility adverb added.


(a) Sheila seduces easily and willingly

(c) The window opened only with great difficulty

(b) Sheila seduces easily and willingly

(d) Asthmatic patients choke easily

75

Fig. 14: A paradigmatic view on the middle construction The paradigmatic opposition with respect to the middle construction has been overlooked in virtually all studies of this construction that I am familiar with (Smith 1978 goes in the direction of observing this ambiguity). In most descriptions of the middle construction, transitives and ergatives are indiscriminately grouped under a general 'transitive' denominator. Interestingly, the Collins Cobuild Dictionary (1987), which tries to make a distinction between transitive and ergative verbs, goes wrong in the other direction by overextending the category of ergative constructions, because it incorrectly labels transitive middle constructions and even agentless passives of transitives as ergative. For instance, they assign the label V-ERG to the usage of fold in the following utterance: (6)

This bed can be easily erected or folded away. (COLLINS)

Clearly, as a result of the incorrect analysis of such a construction as ergative, the number of ergative entries in the dictionary increases drastically, which violates the actual linguistic situation. The incorrectness also surfaces across entries. For example, under the entry erect, the dictionary lists the following example: (7)

This type of chair, easily erected or folded away, is ideal for use in the garden. (COLLINS)

76


This time, Cobuild labels the construction as transitive. The dictionary not only goes wrong in their transitive or ergative labels, it also conflates agentless passives and middle constructions which are two clearly different constructions, as also observed by Fellbaum (1985). The middle construction makes a general statement about the feasibility of the action attributable to particular properties of the Affected. It is indifferent to the Agent that performs the action, i.e. anybody could do it. Agentless passives also defocus the Agent, yet they are tied to a less unspecific Agent. Fellbaum (1985) demonstrates the difference by changes in the adverbs in the two construction types. Consider her examples: (8)a. The butler folded these chairs up easily. (Fellbaum 1985) quickly. competently. clumsily. b. These chairs were folded up easily (by the butler). quickly (by the butler). competently (by the butler). clumsily (by the butler). c. These chairs fold up easily. quickly. *competently. *clumsily. The adverbs competently and clumsily imply facility or ability (or lack thereof) on the part of the Agent that performs the action. Consequently, they cannot occur in the middle construction which is concerned with the facilitating or hindering properties of the Affected. Note that for (8a) and (8b) the way in which the action is carried out is independent of the properties of the chairs. For instance, although they may be of a type difficult to fold, the butler may be handy enough to fold them easily or competently. Given the strong focus on the process-facilitating properties of the Affected, the presence of a "facility adverb" (Vendler 1984) is a typical feature of the middle construction. Fellbaum & Zribi-Hertz (1989:26) assert that le 'besoin de modifier' est une propriété générale des phrases moyennes non impersonnelles [...]. En anglais comme en français, cette contrainte distingue les constructions moyennes et ergatives. [the need for a modifier is a general property of non-impersonal middle constructions [...] In English and in French, this constraint distinguishes middle constructions from ergatives].

The necessity of a modifier is, once more, illustrated by the following examples:


(9)a. b. (10)a. b.

77

This ice cream scoops out easily. (Langacker 1991b) *This ice cream scoops out. Bureaucrats bribe easily. (Keyser & Roeper 1984) *Bureaucrats bribe.

However, while adverbial modifiers are typical, they are only one way of expressing the 'feasibility properties' of the Affected. The limiting case of feasibility is negated doability, as illustrated in: (11) (12)

This dress won't fasten (Fagan 1988) These won't tie. Yes, they tie! (Davidse 1991)

Example (12) shows that the middle construction is also possible in contrastive contexts, reaffirming the properties of the Affected which enhance the doability of the process. In these contexts, contrastive intonation patterns are also functional (cf. also Keyser & Roeper 1984:385, fn. 4). Another way of expressing feasibility is indicated in the following examples: (13) (14)

Stows on floor or shelf. (Fagan 1988) This umbrella folds up in the pocket. (Fellbaum 1985)

The first example describes the way in which a shoe chest is stored, and the information presented in the middle construction is relevant in any discussion of storage. Similarly, the second example states that the umbrella in question can be folded up, which is viable information "since not all umbrellas are folding umbrellas, folding not being the primary function or purpose of an umbrella and hence not an inherent and understood property" (Fellbaum 1985:24). Fellbaum notes that middle constructions occur frequently in advertisements, since the desired effect in this kind of language use is precisely that of "attributing to the product certain properties that can be beneficially exploited by any potential agent" (1985:29). These more subtle forms of feasibility comments may have paved the way for what Davidse has characterized as the ergativization of certain transitive verbs, such as wash, sell or drive. These (originally) transitive verbs can occur in constructions similar to the ergative non-effective (i.e. without any overt feasibility comment, sentence (15c)) and in reflexive constructions focusing on the self-instigatability of the process, as in example (15d): (15)a. b. c. d.

He drove the car fast This car drives easily. This car drove too fast. That car practically drives itself.

[Ac-PROC-Go] [Go-PROC-(Ac)] [ME-PROC] [IS-PROC-ME]

78


Davidse characterizes these verbs as 'blends', combining characteristics of both the transitive and ergative paradigms. They are, in fact, "transitive constellations on their way to becoming ergative ones, but with the transitive principle arguably still dominating" (Davidse 1991:45). The blended nature of these verbs can be represented as an extension from the transitive to the ergative paradigm, as it has diagrammed in Davidse schematic network, reproduced in Chapter 3, section 3.2, as Figure 13. Í agree with Davidse that the transitive middle may be the inroad of the transitive into the ergative, yet want to add that even the ergative (or ergativelike) constructions in which these verbs occur retain a feasibility comment. For instance, when one says this magazine sells, the clause implies that the magazine is of the type that possesses properties that enhance the sales, given that saleability does not constitute, in Fellbaum's terms, an "inherent and understood property" (1985:24) of a product, but is a specific property which sets it apart from other objects. Verbs like drive or sell straddle the border between transitives and ergatives, yet the special properties of the 'Affected' are still relevant to the process. Consequently, the ergative-like uses of these ambivalent verbs will, as a rule, still evoke the notion of contrast, which is reflected in a wide variety of contrastive mechanisms used in these construals. A clear example of this is the use of deictic noun phrases like that car or this magazine, which imply a contrast with other cars and magazines that do not drive or sell as well. Or consider the context of advertisements which, via linguistic as well as non-linguistic means, 'zoom in' on the qualities of the advertised product that differentiate it from product X or Y. The focus on the process enhancing properties of the Affected is generally absent in 'regular' ergative non-effective constructions (e.g. Mary choked) as these properties of the Medium are in fact self-evident, given the inherent coparticipation in the process, even in the passive. This is why ergative middle constructions (cf. Figure 14(c)) do not stand out as much as the transitive middle constructions for which the process-facilitating properties strongly depart from the default reading. So, while I see Davidse's restriction of the middle construction to a "strictly transitive phenomenon" as too absolute (ergative middles with a passive interpretation not being excluded), it is true that, given its special focus, a transitive middle construction is more noticeable and perhaps even prototypical. Yet, as the following section shows, also for transitive middle constructions, a more careful analysis is needed. 4.2.2.2 Lexical constraints on the middle alternation. A serious problem not noted by Davidse is that a middle construction is not possible with just any transitive verb. Keyser & Roeper (1984:383) observe that "not all transitive verbs are equally grammatical in the middle mode":


(16) (17)

79

*French acquires easily. (Keyser & Roeper 1984) *The answer knows/learns easily. (Keyser & Roeper 1984)

but they have to admit that they have no idea why these sentences are excluded. What, then, motivates the grammaticality of some, but not all, transitive middles? The notion of "change of state" is a possible explanation, as offered by Smith (1978), who on this basis draws a distinction between acceptable middles, borderline cases, and ungrammatical ones, illustrated respectively by: (18) The metal anneals easily. (Smith 1978) (19) ?The jewels steal easily. (Smith 1978) (20) *The game watches easily. (Smith 1978) She suggests that "one determining factor is whether the activity referred to by a verb brings about a definite change of state in the object" (1978:104, my emph.). If the 'object' is not clearly affected or not affected at all, the middle is questionable or ungrammatical. The problem with Smith's criterion is how to determine whether the change of state is sufficiently definite, i.e. whether the 'object' is sufficiently affected by the process. Smith's notion of 'definiteness' in fact suffers from the same shortcomings as Levin's notion of 'purity' discussed earlier (Chapter 2, section 2.2.3). Rapoport (1993) also forwards the criterion of change of state as an explanation for the selective occurrence of the middle construction. The inaccuracy of her criterion is indicated by her own need to distinguish between "linguistic" and " 'real world' knowledge" to explain why kick, which in the real world may bring about a change of state, does not tolerate the middle construction, e.g. *Large bears do not kick easily. In the framework of Cognitive Grammar, this distinction is untenable. Linguistic structures are characterized against the background of conceptual domain matrices in which no hard-andfast distinction can be drawn between semantic and encyclopedic knowledge: "there is no essential difference between (linguistic) semantic representation and (general) knowledge representation" (Croft 1993:337). Horn (1980), taking issue with the view proposed by some linguists that reprefixation always involves a change of state, formulates a remarkably parallel criticism: "a judge can reopen a case, but not (felicitously) reclose it, and that while a door can be reopened or relocked, we don't speak of it being reclosed or reshut, despite the fact that closing something affects its state just as much as does opening (or locking) it" (1980:138). I want to argue, once more, that the notion of 'change of state' on which both Smith and Rapoport rely is inaccurate. Firstly, if, as they claim, the degree of affectedness were indeed a determining factor (if not criterial then at least

80


influential), one would expect that kill-expressions, which saliently encode a "definite change of state", readily occur in a middle construction, which, however, they do not. Except for a handful of middle-based OED-attestations for kill, butcher and slaughter (see below), there are no middle constructions with verbs of killing in any of the corpora consulted (on a total of over 15,000 attestations, the few OED-attestations represent a negligible 0.03%). Secondly, and more importantly, the notion of 'change of state' totally ignores the semantics of the middle construction. As emphasized, essential to this construction are the properties of the Affected that influence the doability of the process coded by the verb. What has not been observed in other studies is that very often there is also an implication of 'predestination'. That is, the properties emphasized in the middle construction are often also those for which the entity has been designed in the first place. Fellbaum was quoted earlier as saying out that middles focus on "properties that can be beneficially exploited by any potential agent" and thus often occur in the language of advertisements (examples (6), (7) and (14), reporting on the 'foldability' of bed, chairs and umbrella, are drawn from this type of language use). My notion of 'deliberately designed properties' takes it one step further; although there is a certain relationship between the two, their focus is still different, as will be shown shortly when discussing middle-based constructions with verbs of killing. Instead of resorting to a general notion like 'change of state', I argue that in order for a verb to occur in a middle construction, the semantics of the PROCESS-AFFECTED pairing must be compatible with the notion of process enhancement and/or that of predestination. For some verbs, this condition is not met, and the middle construction is not possible, as for examples (19) and (20). However, with the proper contextual support, creative middle constructions do become possible. For instance, against Rapoport (1993), I claim that it is possible for kick (or almost any other transitive verb) to occur in a middle construction. Suppose that a professional football player comments on a new type of football and utters the following sentence. (21)

Well, I don't know, this new ball doesn't kick very well.

In such a context, the properties of the ball relative to the kicking process become relevant and therefore, the middle construction becomes a valid coding option. The semantic incompatibility between verb and construction explains why verbs of killing do not readily occur in a middle (or middle-based) construction. The reason for this is that it is unusual to think of living beings as possessing properties which facilitate killing them. Interestingly, it is precisely when the


81

notion of predestination (and not that of process-enhancement) comes in that middle constructions may emerge: (22) (23) (24) (25)

As killing cattle or killers (cattle ready for killing) they are inferior to corn-fed stock. (OED on CD-ROM) The very idea of switching our entire herds to milking instead of slaughtering is just crazy. (OED on CD-ROM) His slaughtering stock before the knife would pine. (OED on CD-ROM) Fat butchering cows. (OED on CD-ROM)

In addition to one more attestation with killer, these examples exhaustively represent the middle-based constructions attested.28 They correct Keyser & Roeper's view that "middle verbs are not able to prepose" (1984:387). Noteworthy, all refer to cattle predestined to be slaughtered. Example (23) is especially revealing in this respect, in that it contrasts the predestination of being slaughtered to that of being milked. Notwithstanding these examples, it cannot be denied that it is much harder to devise a context which supports a middle construction for verbs of killing. In addition to the difficulty of conceptualizing the Goal of a kill-event as possessing properties that enhance (or hinder) the killing, the resistance to middle formations is augmented by the strong Agentcentredness of the prototypical verbs of killing. Consider the following usages: (26)

Unarmed people {murder/assassinate/massacre/execute} easily.

A middle interpretation for (26) is not entirely excluded, but the most prominent reading will be the objectless transitive construction, with the unarmed people as Actors, because of their salience in these processes (see Chapter 5 on their Agent-centredness). In short, my claim is that in order to explain why verbs can or cannot occur in the middle construction (or any other construction for that matter) its semantics must be compatible with that of the construction. For some verbs, such as verbs of killing, it is difficult to reconcile the two. However, as has been shown, this is not an absolute issue, since in certain contexts, middle or middle-based constructions become possible. However, given the prototypical incompatibility, such uses can safely be regarded as context-bound extensions and it is highly unlikely that the constructional ambiguities will lead to 'blends' such as shown for sell, wash or drive.

28 Note that all are from the OED on CD-ROM; strikingly, none of them is listed under their respective entries in the book edition.

82


In a more general perspective, the unusual middle construction implies that one must refrain from formulating the constructional possibilities of verbs in absolute terms, as has been common practice (e.g. Levin 1993). Instead, a verb's constructional possibilities depend strongly on the verb's usage. Differently put, there is a strong interaction between lexical and constructional meaning. Such interaction is confirmed by some morphological correlates of the middle construction, viz. the derivation of adjectives in -able and of agentive nominals via the -er suffix (cf. example (22)). Only the former will be discussed here, as the agentive derivation is more elaborately discussed in Chapter 5, section 5.3. 4.2.2,3 English middable verbs. Interestingly, my research on verbs of killing has revealed that, for transitive verbs, there seems to be a correspondence between the acceptability of a middle construction and the adjectival derivation with the suffix -able, which also emphasizes the properties of the Affected in its passive reading, i.e. "having the property to be V-ed". For prototypically ergative verbs, this derivation is unproblematic, given that they by default incorporate the notion that the Affected has the capacities of self-instigating and sustaining the process. Horn (1980: 139) correctly observes that ergative-based derivations are ambiguous between an active and a passive reading (in other words, they preserve the voice syncretism): changeable can mean "that which can change" or "that which can be changed" (cf. also Keyser & Roeper 1984 and Davidse 1991: 106-7). As a rule, the derivation is also possible with transitive verbs, e.g. drinkable, solvable or readable. However, inquiry into the acceptability of the -able derivations for the latter group has revealed a striking correlation with the acceptability of the middle construction. For instance, in the prototypical case the verb kick does not allow a middle construction, and, as a rule, it does not allow the -able derivation either. However, as already pointed out, with the proper contextual support the middle construction is possible and in that context also the -able derivation gains in acceptability. As a counterpart to example (21) with kick in a 'creative' middle construction consider (27). (27)

Well, I don't know, this new ball is not very kickable.

In British English, a kickable penalty is a common expression in football or rugby to refer to penalties that should not present difficulties for the kicker and lead to a score. As with the middle construction, the semantics of the verbs of killing is also incompatible with the -able derivation and formations like murderable, lynchable or assassinatable are awkward. Nevertheless, they do appear, as witnessed by the following examples:

LEXICALLY DETERMINED FLEXIBILITY AND CONSTRAINTS (28) (29) (30) (31) (32) (33)

83

... to kill any man killable by disappointment and sorrow. (OED on CD-ROM) A murderee is a man who is murderable. And a man who is murderable ... desires to be murdered. (OED on CD-ROM) ... the least murderable people in the world, if they could be murdered, then anyone could be murdered. (OED on CD-ROM) Even the angriest demon drivers are reduced to the status of slaughterable black sheep. (OED on CD-ROM) Alexander was a tyrant and therefore in all justice slayable. (OED on CD-ROM) Air-breathing or drownable animal species. (OED on CD-ROM)

In these examples, the -able derivation is fully acceptable because the context focuses on the properties which augment (or justify) the 'killability' of the Affected. While these are relatively unconventional usages, they once more indicate that one must refrain from formulating absolute rules on the acceptabi lity of either the middle construction or the -able derivation. Note that the latter may be more common in co-occurrence with the prefix un- which negates the feasibility. Recently, I came across the following example: (34)

She comes in, green-eye fire hair, sweet stamp of the freshest thing, unkillable lilt ... [LR]

That verbal meaning influences the acceptability of both is nicely illustrated by execute: when used in reference to capital punishment, the verb behaves like the other verbs of killing in its strong tendency to resist the middle construction (cf. example (26)) as well as the -able derivation (e.g. Ίαη executable criminal). However, when used in its more general meaning, "to carry out or perform an action, to carry out an order" (see Chapter 5, section 5.2.4), the formation executable is quite conventional, e.g. a (non-)executable contract. In this con text, it is worth noting that from this general meaning of execute another usage has evolved, viz. the execution of computer commands. Significantly, a file containing one or more such computer commands is called an executable file, its executability being a property which sets it apart from other files. To summarize, with an eye to transitive verbs, of which verbs of killing can be taken as a representative subset, the above analysis has offered an account of the middle construction which does justice to the conditions under which a verb may occur in such a construction as well as semantically related 'morphological' constructions, such as the adjectival derivation with -able. In both cases, the compatibility of lexical and constructional meaning (rather than a general criterion like 'change of state') is essential in explaining when a verb may occur in these constructions. It was further shown that these conditions are not fixed but may be relaxed in the proper context. However, such contexts do not turn

84


verbs of killing into blends of transitivity and ergativity, given the strength of their Agent-centred prototype. In some usages, however, a blend may be at issue, as was illustrated for execute (but then it is not used as a verb of killing). Given the correlation between the semantically related constructions, I have coined the term 'middable verbs', as a cover term for the subset of transitive verbs whose semantic structure is compatible with the middle construction and the -able derivation or other middle-based constructions. Thus, verbs of killing could thus be said to be prototypically non-middable, as their semantic structure is as a rule incompatible with the focus on the process enhancement properties of the Affected. As shown, specific contextual features may overrule the prototypical non-middable nature of these verbs. While I believe that the general lines of this analysis are correct, it is obvious that a more careful and extensive investigation is required. It is only after a more detailed analysis of the middle construction, the -able derivation, and other middle-based constructions has been carried out that the correlation between these (or the absence of it) can be accounted for. A first dimension of further research concerns the -able derivation which is surely worth of in-depth study, as it presents another paradigmatic cut. As pointed out by Keyser & Roeper (1984) and Horn (1980), the derivation is possible with ergatives and transitives, but not with intransitives: transitive (35) a. drinkable solvable readable

ergative intransitive b. alterable c. *sneezable changeable *danceable variable *runnable

As said, the ergative-based derivations preserve the voice syncretism, which is, however, often resolved by the context. For some formations, one of the readings may have a more enduring prominence. For instance, while the derivation (bio)degradable is based on an ergative verb, the conventional reading of the adjective is "that which has the characteristic of decomposing itself by stages". The instigation of process reading ("that which has the characteristic of being decomposed by stages"), although quite plausible, seems to be less foregrounded. The paradigmatic cut is, however, not absolute. As has been shown, verb meaning plays a crucial role in the acceptability of the derivation. Yet Keyser & Roeper also note (but cannot explain) that there are a few exceptional intransitive-based -able derivations, e.g. laughable or dependable. These forms are clearly quite conventional, although the base constructions are to laugh at someone, not *to laugh someone and to depend on someone, not *to depend someone. Another conventional intransitive-based derivation, which is partially


85

related to the lexical field of the present study, is perishable which means "that which can perish", not "*that which can be perished". Faced with such intransitive based formations, as well as complex formations like undryupable, ungetonable-with or unrelyuponable, also Horn (1980) confesses to find the underlying conditions "intractable". Finding a satisfying answer calls for further diachronic and synchronic research, which, however, is beyond the scope of the present work. In the realm of the middle construction, further research is required for the ergative middle construction and its relation to the non-effective. While realizing a passive constellation with the Medium as Affected, a construction like The window opened only with great difficulty, still evokes the notion of coparticipation. However, I suspect that the ergative middle will typically be used when the potential of co-participation is no longer taken for granted, primarily in contexts with negative polarity. Finally, further inquiry into 'middable' verbs is warranted, especially when it comes to including prototypically ergative verbs. In the lexical field of killing, the ergative verbs (e.g. choke, drown, or starve) align with the transitives and have not been attested in 'genuine' ergative middle constructions as depicted in Figure 14(c). Similarly, formations like smotherable or drownable, while possible (cf. example (33) for the latter) appear marked as do their transitive counterparts. The reason is, once again, the semantic incompatibility of verb and construction. Expanding the research to other lexical fields can surely be insightful on all these issues. 4.3 A diachronic perspective on constructional flexibility In the previous section, it was observed that the middle construction may cause typically transitive verbs, like wash, sell or drive to occur in ergative or ergative-like construals. This ergativization substantiates that the two paradigms should not be interpreted as rigid and absolute, but as conceptual categories structured around a prototype and with blurred edges. Certain clauses or clausetypes, like the middle, are situated closer to the borderline of a category and may gradually slide toward the other. Mixed patterns such as for sell and wash, but also other verbs like run or jump have been discussed in the literature, but with the exception of Davidse's work, not in terms of transitivity and ergativity. However, these studies (including Davidse's) have invariably taken a synchronic perspective, i.e. they are concerned with (possibly unusual) alternations as they occur in present-day English. My analysis of verbs of killing, which includes a study of historical material (predominantly drawn from the OED), allows further clarification of such paradigmatic shifts from a diachronic perspective, which is in fact how these changes should be regarded as they generally do not happen over night.

86


Our diachronic exploration first traces the ergativization of intransitive verbs starve and related hunger (section 4.3.1). Next, it will consider the (temporary) ergativization of the transitive SUFFOCATE verbs (section 4.3.2). The last section considers more complex (and particularly verb-bound) paradigm shifts for the verbs throttle, hang and abort. Except for the latter, the discussions open with a condensed summary of synchronic descriptions as found in the linguistic literature. 4.3.1 The ergativization of intransitives Davidse characterizes core intransitives as "constellations that are both 'nonextendible to a Goal' as well as 'non-instigatable' " (1991:113, her emph.). The impossibility of extending to a Goal differentiates them from transitive effectives; the non-instigatability, from ergative effectives. In different yet compatible terms, Smith (1978) indicates that ergative non-efifectives (following Jespersen (1927) she calls them 'change verbs') have the features [+Independent Activity] and [+External Control], whereas intransitives are marked [+Independent Activity] and [—External Control]. Crucially involved is the latter feature which indicates that the instigation cannot be relinquished to an external cause. Similarly, Fellbaum & Zribi-Hertz (1989:34) distinguish 'Verbes ergatifs" from "verbes inaccusatifs non ergatifs" on the grounds that the former, but not the latter, may open up to include the Instigator. A change from intransitive to ergative, which is what concerns us here, thus implies a move from a fully autonomous process to a semi-autonomous one, i.e. a process that has become conceivable as instigatable. Davidse observes that ergativization is quite common with intransitive processes involving technical devices, which can be conceptualized as having an energy source of their own: (36)a. b. (37)a. b.

John switches off the machine. (Davidse 1991) The machine switches off automatically. The motor is running. (Davidse 1991) She is running the motor.

Elaborating Davidse's insights, Geyskens (1997) provides a corpus-based study of the ergativization of prototypically intransitive manner-of-motion verbs with animate Actor, such as run, jump, and march as well as some other verbs (e.g. bleed and work), correcting in the process some inaccuracies in the linguistic literature on this phenomenon (see e.g. Cruse 1972, Smith 1978, Pinker 1989, Levin 1993, Levin & Rappaport 1995). Within the field of killing, there is one verb, starve, whose lexical and grammatical evolution is particularly interesting in this respect. As will be discussed shortly, the semantic and constructional extension have drastically changed the original situation for this verb. Interestingly, in the course of this


87

extension process, the verb came conceptually close to the verbs hunger and thirst which will be briefly discussed in the second part of this description. 4.3.1.1 Starve. In the Old English literature, the verb starve covers the same meanings as present-day die, as the following examples show: (38) (39)

Annanias and Saphiran ... steorfende afeollon (OED, c. 1000) "Annanias and Saphiran ... fell dying" Se man .. .Þe nan ne heafde stærf of hungor (OED, 1124) "The man ... that didn't have to die from hunger"

In this original usage starve reveals its formal and semantic affinity with other (West-)Germanic languages (e.g. Old Frisian sterva, Old Saxon sterban, Middle Dutch sterven, and Old High German sterban). Given that there is no attestation of the verb die in the OE literature, it is uncertain whether or not this verb existed. Certain is that, under theinfluence of the Scandinavian verb dejyan, the verb die comes into general use in the 12th century. The verb covers the same ground as starve, both in its lexical and constructional properties. Although there might have been subtle semantic differences between the two verbs, the apparent near-synonymy and interchangeability may have triggered starve to undergo a process of semantic specialization, whereas die has remained fairly constant in its usages (strikingly, none of the many meanings listed in the OED is marked as obsolete). More specifically, out of its relatively wide range of types of dying, starve starts to profile one particular kind, referring to a lingering death caused by hunger (examples (40) and (41) below), cold (example (42)), grief, or slow disease. Also common were metaphorical extensions of this usage, such as expressing extreme suffering, example (43), or pining away with desire, example (44). The OED-data indicate that this new profile manifests itself by the end of the 16th and in the course of the 17th century: (40) (41) (42) (43) (44)

... the prisoners ... that were steruing for famine (OED, 1618) An Animal that starves of Hunger, dies feverish and delirious (OED, 1735) So out of doores go they with his clothes ... and left Nicholas Nouive starving and quaking in that dog-hole. (OED, 1602) The weather grows cold ... I'll go rise, for my hands are starving while I write in bed. (OED, 1710) Though our soules doe sterve For want of Knowledge, we doe litle care. (OED, 1616)

The addition of'for, of or for want of occurs regularly with starve to emphasize the shortcomings from which the victim suffers. In contemporary English, the

88


for-complement is not necessary anymore, as in literal usage there is only one possible shortcoming when starve is used, viz. that of nutrition. This means that from the 17th century up to the present, starve has further specialized to mean "to die for, or suffer extremely from, lack of food". However, when used metaphorically, a for-complement specifying what is yearned for is common, e.g. a company starving for cash. Interestingly, the lexical shift is accompanied by a constructional change, as the verb starts to occur in a two-participant construction (the earliest attestation listed in the OED is from 1529), in literal as well as in metaphorical usage: (45) (46) (47)

He ... also sterued them for honger and cold, so that many died. (OED, 1560) Who abuseth his cattle and starues them for meat (OED, 1573) In steede of consuming and staruing your euil, you giue it nourishment (OED, 1581)

This is a clear instance of ergativization: in its original meaning "to die", the process encoded by starve was clearly non-instigatable. However, its evolution towards the conceptualization of a lingering death resulting from deprivation created the semantic conditions for the inclusion of an Instigator, a possible external controller depriving the Affected of vital needs. In other words, the verb has come to denote a process of death conceivable as instigatable. While initially this ergativization may have been an extended usage, the ergative pattern has established itself firmly in the course of time and the verb has lost its hook-up with the intransitive to become a full-fledged ergative verb. In contemporary English, the non-effective construction, like In those days Poland was starving, syncretizes both the notion of self-instigated and externally instigated, as is typical of ergatives. This syncretism gives rise to ambiguities in the interpretation of past participles, a construction which, next to the non-effective, is quite common with starve (as is also typical of ergative verbs). For example, it remains unclear whether in (48)

"Married ... With Children" is the story of the Bundys — snotty younger brother Bud, trampy older sister Kelly, sex-starved spandex-clad wife Peg, and terminally boorish husband Al (WSJ)

spandex-clad Peg is simply always hungry for sex or whether she is deprived of it by her boorish husband Al. In fact, both readings are entertained simultaneously. As said, for metaphorical uses, what one is being deprived of is commonly expressed in an of-complement or a for-complement. In itself, the choice of preposition does not resolve the question of instigation, as shown by the following examples of externally instigated processes:


(49) (50)

89

The Argentine Navy has been starved of spare parts for its British-built ships ... (WSJ) ... stores "have been basically starved for merchandise" because of supplier worries about payment ... (WSJ)

A possible relic from the past is that, perhaps more than for other ergative verbs, the caused process (the starvation) has a relatively high degree of independence from the Instigation. This may, however, equally be attributed to the verb's lexical load, i.e. the reference to a slow process of undernourishment, eventually leading to death. Moreover, as is also typical of an ergative conception of events, there is a low salience of goal-achievement, which contrasts sharply with the foregrounding of this property in the case of prototypically transitive verbs, like kill or murder (see Chapter 5). The exclusion of the endpoint of the process is what motivates the usage of starve in reference to the state of being very hungry (usually but not necessarily because of not having eaten for some time) and not to an actual dying process due to lack of nutrition. In fact, for starve, the death of the Affected often seems to be excluded in the conceptualization altogether. Just recently, I heard an interviewee's utterance be completed by a BBC reporter (in voice-over) in the following way: INT. REP.

... mother seals don't have enough energy to produce milk and puppies are starving and dying.

For the reporter, the two events are clearly different. In other words, the strong focus strongly on the Medium's activity not only leads to excluding the instigation but also the end-point of the process. The salience of the starving participant is also reflected in the possibility of Medium profiling starver, as attested in the OED, which is uncommon (see Chapter 5, section 5.3). 4.3.1.2 Hunger and thirst. The new lexical extensions of starve make it encroach on the conceptual space of intransitive hunger. Interestingly, when starve starts to occur in an ergative effective construction, also hunger appears to undergo a similar process of ergativization: (51) (52)

The Prince ... came himself To hunger Middleburgh (OED, 1575) Theire pasture will hunger our beasts ... (OED, 1641)

In other words, both verbs become strongly aligned lexically as well as constructionally. There is an interesting formal reflex of their relationship in the formation hungerstarve that was common for some time, and displayed ergative characteristics as well, e.g. To save us from hunger starving (OED, 1592). In

90


later uses, the modifying hunger was dropped as it became part and parcel of the semantic profile of starve itself. Unlike with starve, the ergative pattern has not entrenched for hunger and much of the lexical proximity has also been lost, since the verb is only used metaphorically in present-day English. It is now a verb of desire (Levin 1993), as are verbs such as long, yearn or crave. What triggered this change is hard to say with absolute certainty, yet undoubtedly influential has been that, as with starve, this verb was commonly used hyperbolically, referring to the state of being extremely hungry, a conceptualization which excludes the end-point of the process. The same evolution to a verb of desire can be observed for thirst which in its original intransitive usage referred to the state of being thirsty. Remarkably, however, this verb has not been attested in ergative uses. Why this is so remains at this point unclear. For the sake of completeness, it can be mentioned that in metaphorical usage both hunger and thirst have been attested in Ranged constructions, in which the thing being desired surfaces as the syntactic object: (53)

Blessid be Þei Þat hungren and thirstun riztwitnesse (OED, 1382) "Blessed are they that hunger and thirst for righteousness"

In such constructions, the second participant does not realize the feature of affectedness, and hence the clause is not effective. We will not go into full detail on the complexities of pseudo-effective constructions (see, however, Davidse 1991:115-140), but a brief note is in order. In pseudo-effective constructions, participant status is conferred onto an element which is only indirectly related to the process. Davidse observes that this brings the circumstantial element into more direct interaction with the process, increasing the notion of affectedness (cf. also Langacker 1991b:303). Yet, the process, as a rule, is never truly effective, as shown by the impossibility or markedness of a passive (e.g. *The city was prowled, ?The lake was rowed).29 Whether such more direct interaction is also true for Ranged constructions with thirst and hunger is impossible to tell from the OED data, although an occasional contextual element seems to corroborate the implication of an increased strength of desire. Consider, for example, the adverb greedily in

29

This is not entirely accurate, as the Range category is scalar, and some constructions come close to encoding affectedness, as in They climbed the mountain (versus They climbed on the mountain). The higher degree of effectiveness is shown by the (increased) acceptability of a passive: The mountain has been climbed by many tourists.


91

(54) Þi mowth hungreth gredyly delycaycyes ... Þin erys hungryn gredyly newe tydynges (OED, 1440) "your mouth greedily hungers for delicacies ... your ears hunger greedily for new tidings" The OED does not list any passive constructions, but given the relatively low number of attestations, caution is warranted. In any case, it should be noted that a Ranged construction (i.e. with a pseudo-Goal) is restricted to intransitive processes and are not possible with ergatives, e.g. *They starved a gruesome starvation or *They starved cash (with cash being the thing craved for). This may explain why the Ranged construction continued to occur with thirst until the 18th century but not with hunger for which the OED lists no attestation after 1563. By that time, hunger had started to behave like an ergative verb. In other words, from the 16th until the 19th century hunger and starve were lexically and constructionally similar, whereas thirst continued to code an intransitive process. What triggered the later lexical and constructional evolution to the present-day situation requires more elaborate research than we can afford to do here. Suffice it to reiterate that in present-day English hunger and thirst are clearly intransitive verbs of desire, whereas starve has entrenched as a fullfledged ergative verb. 4.3.1.3 Ergativization as linguistic creativity. The above descriptions of starve and hunger seem to give credence to Halliday's claim that "the coming of [the ergative] pattern to predominance in the system of modern English is one of a number of related developments that have been taking place in the language over the past five hundred years or more" (1985:146). Note that an ergativization may show up unexpectedly, as in the following ergativized use of die: (55)

... in such gradual sensible death ... God dies away in us, as I may say, all human satisfaction, in order to subdue his poor creatures to himself. (OED, 1748)

The phrase "as I may say" indicates that the author is well aware of stretching the constructional potential of die, yet the logic behind his coding is straightforward: to ensure subduction, God (certainly in those days the Instigator par excellence) directly causes (by impinging "such gradual sensible death") all human satisfaction to die away. While such an (idiosyncratic) ergativization of die has not entrenched, the potential for ergativization is still there. I recall a colleague reporting how, when they had found a hedgehog in the garden, her youngest daughter said to her father in an over-ergativizing mode "Don't die it" (Kristin Davidse, pc). Such an ergative usage of die is not sanctioned in

92


'proper' English, yet in addition to example (55) it illustrates how language users exploit the paradigmatic oppositions. Unexpected and 'creative' instances like these reveal, to put it in Hallidayan terms, 'the system behind the text'. To sum up, the evolution of starve is a clear instance of ergativization of an originally intransitive verb. Because of its lexical and constructional evolution, the verb has distinguished itself from its allies in other Germanic languages (e.g. Dutch sterven or German sterben) which have continued the original meaning and construction types. As a refinement to Davidse's framework, the description has offered another indication of the importance of lexical meaning in the description of transitivity and ergativity. Our analysis has further refined Davidse's work in yet two other ways. First, Davidse's description of the ergativization of intransitives (1991:115) is restricted to 'controlled' intransitive constellations and is said not to occur with intransitives that have an Actor who is not in control, as with faint, stumble, blush or die. The evolution of starve has provided further evidence that 'uncontrolled' intransitives may be subject to ergativization, as has also been pointed out, for instance, by Smith (1978) commenting on the sentence The nurse burped the baby or by Levin & Rappaport (1995) commenting on The doctor bled the patient. As will be elaborated further in the final conclusions to this book, a problem with the Levin & Rappaport's views is that they see extensions such as these as "spurious causativization" and thus fail to see the semantics of the system behind them. Secondly, our description has shown that what originates as a constructional extension may become a well-entrenched pattern, unlike many of the ergativized uses mentioned in the literature, e.g. Davidse (1991), Levin (1993), Levin & Rappaport (1995) or Geyskens (1997). The latter corpus-based study confirms the intransitive prototype of manner of motion verbs statistically, as intransitive uses of these verbs range from 90% to 99%. Clearly, in the case of starve, the specific lexical evolution can be held responsible for the conceptualization of a non-volitional intransitive process as instigatable and for the subsequent entrenchment of this ergative pattern, since the latter harmonized well with the new conceptual profile of the verb. Such harmonization of lexical and constructional profile can also be seen in the ergativization of originally transitive SUFFOCATE verbs discussed next. 4.3.2 The ergativization of transitives That ergativization is common in modern English has not gone unnoticed in the literature. For example, Keyser and Roeper (1984) observe the productivity of the ergative pattern in modern scientific and bureaucratic English, as, for instance, reflected in the ergativization of processes expressed by verbs ending in -ize (oxidize, federalize, etc.). They immediately add, however, that not all


93

-ize forms ergativize, but that it involves a combination of -ize and the properties base form. The ergativization of intransitive starve and hunger confirms the emergence of the ergative pattern. Another particularly interesting evolution in the field of killing is the ergativization of the transitive SUFFOCATE verbs. The OED data on the SUFFOCATE verbs indicate that their evolution is characterized by ample lexical and constructional overlapping. Although the two are strongly intertwined, we will in this section only deal with the constructional convergence. A discussion of the lexical convergence is postponed to Chapter 6. Even in present-day English, the SUFFOCATE verbs show a marked preference for ergativity, but at some point in the history of English, roughly from late 16th century onwards, all SUFFOCATE verbs occur in ergative constructions, including the originally transitive verbs strangle, throttle, stifle, and smother. Consider the following examples: (56) (57) (58) (59)

He could not lie down ... he should strangle (OED, 1662) I smoother with the smoake ... (OED, 1648) ... we are euery houre in danger of stifling (OED, 1594) She dyed without any Fever ... without throatling. (OED, 1687)

The ergative pattern, which occurs in literal as well as metaphorical usage, continues until the early 20th century, and for smother and stifle ergative constructions have been attested in contemporary English as well, witness examples (63) and (64): (60) (61) (62) (63) (64)

She fell into the pond yesterday ... She nearly strangled .... (NOV) The child throttled and died in my arms (OED, 1909). ... he would not go into this hole because he was too hot and would smother (OED, 1895) The sharp jab of cold steel replaced the smothering pain ... (CASS) The garrison stifles in rock heat (OED, 1966)

In addition to the ergativization of intransitive starve and hunger, the ergativization of all SUFFOCATE verbs may be taken as further evidence in support of Halliday's claim regarding the productivity of this process in Modern English. As said, the constructional convergence co-occurred with and was probably due to a high degree of lexical overlapping. As Chapter 6 will substantiate, the verbs occurred in seemingly identical contexts in identical constructions. The description in Chapter 6 will further show the historical and contemporary ergative predilection of the SUFFOCATE verbs to be experientially motivated. As with starve, the (subtly differing) conceptualizations designated

94


by the SUFFOCATE verbs all entail a strong profile on the activity of the Medium, often excluding from the conception both instigation and end-point of the process. In other words, our analyses refine Halliday's quite general claim by showing that the ergativization of the intransitive verbs starve and hunger and the transitive SUFFOCATE verbs are grounded in our experience of reality. The situation is quite different for the MURDER verbs (e.g. kill, murder or slaughter) which have a built-in focus on the Agent. None of these verbs has been attested in ergative constructions, nor is it likely that they ever will. For example, even if you tried hard, it would be difficult to convince your audience that Tony murdered is to be understood as an ergativized construction with Tony realizing the role of the Medium. Unlike starve, the ergative pattern has not entrenched for the SUFFOCATE verbs in their evolution to present-day English nor has it converted them into blends of transitivity and ergativity. The originally transitive strangle, throttle, smother and stifle have 'retransitivized', although the situation is perhaps somewhat more complex for the latter two, as suggested by examples (63) and (64) instantiating an ergative pattern. A more elaborate clarification of how these usages fit into the semantic networks of both verbs is offered in Chapter 6. At this point it should be sufficiently clear that the global ergativization of the SUFFOCATE verbs presents another illustration of the influence of a verb's lexical structure on its constructional potential. Some examples of paradigmatic changes more severely constrained by a verb's usage are discussed in the next section. 4.3.3 Complex patterns ofparadigm shifts In general, the paradigmatic shifts discussed so far have led to clear results: either the outcome is an entrenchment of the new pattern, as with starve, or the extensions have been abandoned in favour of a further entrenchment of the original pattern, as with hunger or the SUFFOCATE verbs (disregarding for now the less common ergative uses of stifle and smother illustrated above). However, the paradigmatic changes can lead to more complex results, viz. blends of the two paradigms with one and the same verb. Such constructional blends have, in fact, already been illustrated in the earlier discussion of the verbs wash, sell and drive. Within the field of killing, some other illustrations can be offered, viz. the ergativization of one particular usage of throttle, the transitivization of hang in its usage as a verb of killing, and the transitive/ergativity split characterizing abort. 4.3.3.1 Throttle an engine. Apart from the 'general' ergativization of throttle mentioned in the preceding section, there is one particular extension of the verb


95

which has paved the way for a process of ergativization. The extension maps the image of (gradually) narrowing the windpipe onto that of controlling gasses and/or fluids in tubular parts of engines. Not incidentally, it is first attested in the 19th century when steam engines and other types of engines were being developed and experimented with. A typical usage is the following: (65)

It would be better to use the steam expansively, rather than to throttle it by means of the regulator. (OED, 1875)

What motivates this extended use is the similarity between the tube through which the steam escapes and the human windpipe: the throttling reduces or stops the emission of steam as the constricting of the windpipe stops in- and exhalation. In this sentence, the act of throttling contrasts with the phrase to use the steam expansively. Via generalization, this usage will bleach into one which merely refers the notion of regulatory control. Eventually, this will lead to a reversal of the original metaphorical construal: instead of quantitative decrease, throttle can nowadays also refer to quantitative /«crease. For instance, running a motor full throttle means running it at maximum power by maximal injection of fuel. Alternatively, one can throttle down an engine to idle, meaning that it will run at the lowest possible speed. Throttle has thus come to mean "to control the supply of fuel into an engine" and other elements in the context will specify in which direction the fuel supply goes (e.g. full vs. down to idle). In fact, the expression full throttle designates a conceptualization that is antonymical to the one it started from, as in the original conception full throttle indicated complete closure of the regulating valve, halting the flow within the tube. A factor which may have been influential in this development is the metonymical shift from throttle [GAS] to throttle [ENGINE].30 Hence, a more accurate paraphrase of the verbal usage would be "to use the throttle-valve (or something like it) to regulate the quantity of fuel injected into the engine". Most interestingly, the profiling of the engine has laid the basis for a process of ergativization of the type mentioned earlier, viz. involving a technical device or engine which is seen as having an 'energy source' of its own which directly operates in the process. The following attestations with throttle (both as noun and verb) may serve to elucidate such an ergativization as well as the resulting mixed character of the item. All sentences are taken from the same scene of a science-fiction story (involving a self-made spaceship with a diesel engine); the original discourse order has been preserved:

See Chapter 3, section 3.1, on the mechanisms of metaphor and metonymy.

96

LEXICAL PERSPECTIVES ON TRANSITIVITY AND ERGATIVITY (66)

... ... ... ... ...

the engine roared at full throttle ... he throttled down the diesel and pushed in the clutch ... You should apply the throttle a little more gently ... The diesel again throttled down to an idle ... Lester throttled down and pushed in the clutch. (CASS)

The ergativized usage has been marked in italics. Note, however, the objectless transitive construction in the last sentence. This signals the blended nature of throttle, since for ergative constructions, omission of the Medium is generally impossible (*John starved cannot be interpreted as an ΙΝSTIGATOR-ΡRΟCΕSS constellation). This specific ergativization of throttle is independent of the 'global' ergativization discussed in the preceding section which, as will be recalled, no longer occurs with throttle. The ergative the diesel throttled down stands out as a separate extension, which in fact has become so entrenched and semantically remote from its source that some speakers regard this usage as the 'actual' meaning of the verb and consider it no longer a verb of killing.31 4.3.3.2 Hang. Another example of a paradigmatic change constrained by usage is the verb hang. This verb, when used in its general meaning, viz. "to suspend vertically from above so that the lower end is free", activates the ergative model: (67)a. She is hanging the picture on the wall. b. The picture hangs in the living room. However, when referring to a killing ("suspension by the neck until dead"), the verb evokes the transitive paradigm only. In Modern English, the difference even finds its formal reflex in the different forms for the past participle, hung versus hanged. As Jespersen (1927) points out, the difference goes back to an OE difference between the strong verb (hon), which was transitive and the weak form hangian which was intransitive. However, as far as I can judge from the OED attestations, there was early confusion of the two in Middle and Early Modern English. The formal opposition in present-day English thus seems to be a modern reshuffling, although Jespersen mentions that also in the meaning of "put to death by hanging" the strong form hung is gaining ground. In any case, in this use the verb is clearly transitive, which lines up with the overall transitive orientation of the field of killing (see Chapter 3, section 3.2). 4.3.3.3 Abort. A final example of a diachronic evolution leading to a paradigmatically complex category is the polysemic cluster abort. We can be 31

One of my proof-readers (native British) made a comment in this vein.


97

brief here, as its lexical and paradigmatic fluctuation will be treated in Chapter 7, after the semantics of the transitive and ergative paradigms have been more fully described. In a nutshell, the evolution boils down to the ergativization of an originally intransitive verb, which can be regarded as an exponent of the general ergativization to which also the SUFFOCATE verbs were subject. An additional layer of complexity for present-day abort is that is has partially transitivized. So far, the latter has affected the literal uses only, resulting in a transitive/ergative split which parallels the literal/metaphorical distinction. However, the transitivization is still an on-going process and, although it is impossible to make firm predictions, individual comments made to me by native speakers seem to indicate that the transitivization is extending to metaphorical uses as well, possibly leading to a fully transitivized category (see, however, some further complications discussed in Chapter 7). The case of abort is in fact a good illustration of how our conceptualization of the world can have a direct reflection in grammatical patterns. In fact, all the paradigmatic changes discussed in this section (as well as other phenomena discussed in the case studies, e.g. the transitivization of choke when combining with particles like out or off) evidence that the import of the ergative and transitive paradigms as described here is cognitively real and is at issue in our experience of actions and events. 4.4 Conclusions The synchronic and diachronic description presented in this chapter have given a first assessment of the complexities involved in constructional variability (or the absence of it) with verbs of killing. The lexical-paradigmatic description of the middle construction has emphasized the idea that this construction (or any other construction for that matter) is meaningful in and by itself. The meaning of the middle construction was characterized as implying a profile on the properties of the Affected that facilitate the process or predestine it to be submitted to the process. An Agent is implied, yet remains quite schematic as the properties are such that any Agent could partake successfully in the process. The description has further revealed the paradigmatic complexity of the middle construction generally overlooked in studies of this construction and has offered a coherent account of its acceptability which does justice to the semantics of the construction as well as that of the instantiating verbs. The diachronic description of some paradigmatic changes has evidenced the non-rigidity of the transitive and ergative paradigms. Instead, it has been shown how verbs may temporarily or permanently shift their constructional prototype. In all cases, lexical extension could be identified as paving the way for these constructional extensions. The complex interaction of the two paradigms is

98


described extensively, both from diachronic and synchronic perspective, in the analysis of the verb abort. To fully understand the semantic impact of this transitive/ergative interplay, we will first explore the semantics of each paradigm separately in the two following chapters. Chapter 5 clarifies the Agentcentredness of the transitive paradigm, Chapter 6, the Medium-centredness of the ergative paradigm.

CHAPTER 5 THE AGENT-CENTREDNESS OF THE TRANSITIVE PARADIGM

5.1 The experiential basis of agentivity Like many other West-European languages, English can be classified as a morphologically nominative/accusative language: in the active voice, the intransitive and transitive Agent appear as Subject, i.e. in the nominative. The predominance of the transitive system in many West-European languages may be due to the deeply engrained anthropocentric world view, in which man aspires to control the world (cf. Wierzbicka 1991, 1992). This notion of control is maximally activated in a transitive conception of an event in which a volitional being wills an action and effects it to impinge on another entity. But what then with inanimate entities that may surface as the subject of transitive verbs? Are they to be regarded as Agents? What about accidental actions which have an effect that was not desired by the Agent? Before the Agent-centredness of the MURDER verbs can be investigated, these more general questions concerning the nature of agentivity must be answered. Inspired by a number of recent publications, including Cruse (1973), De Caluwe (1995), DeLancey (1982, 1983, 1984a,b, 1985, 1986, 1987, 1991), Schlesinger (1989) and Nishimura (1993), I develop a cognitive definition of the Agent role. More specifically, I argue that the notion AGENT represents a prototype-based category whose structure is experientially motivated, thus correcting in the process some inaccuracies of common views on semantic roles in general and the Agent role in particular. It should be stressed that the redefinition of Agent pertains to the more schematic category Agent which generalizes over the two paradigmatically determined instantiations, the transitive Actor and the ergative Instigator. Clearly, on a more delicate level, the two types of Agent cannot be conflated, but for the more general discussion, the distinction may be ignored. As already emphasized by the scholars cited above, the notion of control is central to the notion of agentivity and thus also to the transitive model. DeLancey and Nishimura correctly identify the prototypical transitive Agent as a volitional actor in full control of the event (from act of will to achievement of the goal) and who is thus also fully responsible for it. Non-volitional Agents, they argue, are less typical and so are inanimate Agents; both are extensions

100


from the prototype, as they suspend the notion of intentionality and/or animacy. The following attestations from the WSJ corpus illustrate these different types of Agent. For each type of Agent, both a transitive (a) and an ergative (b) construction is given: volitional Agent (l)a. She suspects that he's decided to kill himself. (WSJ) b. ... suddenly he'd ... grab him by the throat and choke him ... (WSJ) non-volitional Agent (2)a. If we killed any tortoises, we ... didn't purposely do it (WSJ) b. ... [a child nearly] drowning itself ... with each new experiment (CASS) inanimate Agent ('Instrument') (3)a. ... tiny amounts of DDT killed flies ... (WSJ) b. ... a heavy, colorless gas that ... can suffocate humans. (WSJ) inanimate Agent ('Force') (4)a. At least 16 people were killed by the storm in the Philippines ...(WSJ) b. ... construction choked traffic at the intersection ... (WSJ) Characterizing the so-called Instruments and Forces as Agents goes against some of the more traditional views on case roles, especially as defined in (early) case grammar (cf. Fillmore 1968, 1972, 1977 or Starosta 1978). The problem with these views is that they often rely too directly on the extra-linguistic world to characterize semantic roles. According to these views the DDT in example (3a) would be assigned the role of Instrument, since 'in the world out there', it is manipulated by humans to destroy pests. The underlying structure of (3a) would thus be one in which these humans are the Agents, the poison is the Instrument, and the pests are the Patients. This underlying structure may be coded in a variety of ways (e.g. During World War I, soldiers killed flies and mosquitoes with DDT), yet it is claimed that such alternations (different surface forms) do not alter the thematic relations of the deep structure. Cognitive grammar rejects the role-preservation principle, since one of its basic assumptions is that a difference in construal always entails a difference in meaning. Alternating constructions like Farmers kill flies and mosquitoes with DDT and DDT kills flies and mosquitoes are semantically distinct, since they impose a different image on the event. Although some people may be uncomfortable with the idea of inanimate agents, in sentences like The window was broken by a stone or Poison killed Jane, the subject nominals express entities that qualify as agents "understood as primarily responsible for the broken window and Jane's death, respectively; they are construed as doing the breaking

THE AGENT-CENTREDNESS OF THE TRANSITIVE PARADIGM

101

and the killing" (Nishimura 1993:496; his emph.). Paraphrases in which poison and stone surface in a with-phrase are of course possible, but construe the inanimate entities stone and poison as subordinate to some other entities (most likely, humans) that did the breaking and the killing. In DeLancey's words, "an instrument [...] makes no independent contribution to the event—it functions only as an extension of the agent's will". (1984b:207). However, the 'promotion' to Agent of an entity which pre-linguistically clearly counts as Instrument is not always possible: (5)a. b. (6)a. b.

The terrorist killed him with a bomb.(Nishimura 1993) A bomb killed him. The terrorist killed him with a stick. (Nishimura 1993) *A stick killed him.

Some linguists (e.g. Wojcik 1979, Marantz 1984) have tried to account for this phenomenon by distinguishing between "enabling/facilitating" instruments and "non-enabling/intermediary" instruments. The latter instruments can perform the action autonomously; the former cannot, and therefore cannot occur in subject position. This account is partially correct in observing a semantic difference, yet the distinction should be seen as one reflecting a difference of conceptualization, rather than an ontological one. The cognitive account, which regards bomb and stick in the (a)-sentences as instruments but as inanimate agents in the (b)-sentences, explains the acceptability of (5a) and unacceptability of (6b) by the possibility of "independent action" attributable to the entity (DeLancey 1984b:203). A bomb is sufficiently capable of independent action, the act of exploding (cf. The bomb exploded), which facilitates construing it as agent. However, regarding a stick as agent is less plausible and requires strong contextual support. Yet contrary to what is often assumed, possible contexts for 'inactive instruments' need not necessarily be restricted to wild fantasies about magical sticks, axes, or knives. A typical context is one in which direct responsibility is shifted to some other entity (cf. Deane & Wheeler 1984:97, fn.l).32 DeLancey, Schlesinger and Nishimura observe that the context often helps to make intelligible otherwise contrivedsounding sentences: (7)

32

The poison which she had slipped into his coffee killed him. (Nishimura 1993)

I have noted that our children regularly exploit the shift of responsibility to outsmart their parents, e.g. in a sentence like It's not me who is splashing, it's this duck (sitting in the bathtub and splashing vehemently). It would be interesting to examine such shifts more systematically.

102


(8) (9) (10)

A bullet fired from his own gun killed Tom by penetrating his heart. (Nishimura 1993) The axe fell off the shelf and broke the window. (DeLancey 1984b) The lawn mower threw up a stone and broke the window. (DeLancey 1984b)

Clearly, it is most relevant to consider the larger context of examples in order to arrive at fine-grained distinctions that do justice to linguistic data. My corpus material further indicates that a speaker's choice of a given construal is not an isolated phenomenon, but the result of a coding decision in which many factors are intertwined. A given coding may be experientially conditioned or may be the result of a conscious choice. Consider, for instance, the following examples from my corpus: (11) (12)

... a fiery explosion ... killed at least 160 workers. (WSJ) Trooper casualties numbered 25 — mostly killed by stray bullets from their own rifles or shrapnel from the artillery. (WSJ)

In the context of the accidental explosion, construing the event in agent as example (11) seems a logical choice, since it is difficult to trace the chain of causation further back to a more ultimate cause. For example (12), reporting on the Battle of Wounded Knee, ideological arguments may have played a part in the coding, as the writer sides with the Indians. He restates the battle as a "massacre of mostly unarmed men, women and children" (my emph.). When he mentions the Indian casualties, the troopers are seen as the agents; the next sentence, example (12), contrastively takes an inanimate agent which may be interpreted to further emphasize the lack of violence on behalf of the Indians. Another motivation for casting the 'instrument' into the agent role comes to the fore when the action or process described in the sentence is crucially dependent on some property of that instrument (Schlesinger 1989:191). Consider the larger context of example (3) above: Mueller had been engaged in a systematic search for a new insecticide with very specific properties. It must be very toxic to insects, long lasting in its effects, non-toxic to man and to plants, and cheap and easy to produce. ... he worked for several years through a whole series of new compounds. Several proved to be promising insecticides, but DDT was the best. ... He found that tiny amounts of DDT killed flies, aphids, cockchafers, Colorado beetles and many others. (WSJ) Given the contextual focus on the characteristics of the poison which guarantee its efficiency, it is most natural that it is construed as Agent. Clearly, the


103

instrument's properties need not always increase the feasibility of the action, but may instead hinder it: (13)

The diluted chemical puts stress on weeds without killing them, greatly increasing their susceptibility to the bacteria. (WSJ)

An interesting experientially-bound influence on the coding of the instrument as Agent is pointed out by Ryder (1991b). She observes that as the actions of agent and instrument become more separated in time, the agent's action may be construed as outside the scope of the event, leaving the instrument as the most agent-like participant remaining. Slow working poisons and time bombs are prime examples in this respect. The temporal separation of the agent's action and the instrument's action are in fact a specific instance in which the instrument is more readily envisaged as capable of independent action. The greater temporal distance may (but need not) trigger codings with less agentive constructions, like an analytical causative or a 'circumstantial construction', like e.g. John died in the bomb explosion (I will return to the latter type of constructions when discussing kill). Further, my data for kill, the verb showing the greatest variety of Agents, have revealed that there is a strong correlation between the occurrence of an inanimate Actor and the type of Goal. In general, there is a certain agreement of rank between Actor and Goal when they are interpreted in terms of Silverstein's empathy hierarchy. This hierarchy reflects, as Langacker puts it, "an egocentric assessment of the various types of entities that populate the world and ranks them according to their potential to attract our empathy."(1991b:306).33 Adapted to my coding scheme, and leaving out speaker and hearer as the first two entities, the hierarchy can be represented as follows: human>human-like>animal>organism>physical object>abstract object 'Human-like' stands for metonymical extensions to governments, agencies, countries, and the like. Now, in the case of kill, patients higher in the hierarchy occur more frequently with an Actor of comparable rank (e.g. human(-like) Actor and human Goal); Goals situated at the lower end of the hierarchy also prefer Actors of lower rank. The correlation is conceptually well-motivated. As already pointed out, the conceptualization of an event consists of a series of temporally contiguous segments which together make up an event-chain. Within such an action chain, we may group together different units, which Ryder

See also Deane (1987) for a detailed description.

104


(1991b) conveniently terms 'episodes'.34 Despite a certain amount of natural clustering among events in a chain, what is construed as an episode can vary greatly. When it comes to the destruction of plants, pests or bacteria, the human agents that are usually behind the destruction will mostly fall outside the episode, leaving the substance that directly acts upon the Goal as the most agent-like participant. This accounts for the correlation between low-level Actors and low-level Goals. Language users must make choices and decide at which 'level of episode' they want to express a given event. Observe, however, that the choice is an either-or affair, since one cannot conflate two episodes into one. Consider the following examples: (14)a. scientists discovered an antibody that... kills AlDS-infected cells (WSJ)

b. scientists succeeded in killing AIDS-infected cells. (15)a. *The antibody and scientists killed AlDS-infected cells. b. *AIDS-infected cells are killed by scientists by the antibody. Both (14a) and (14b) are correct, coding as Agent, the 'manipulator' and the 'instrument' respectively, but (15) shows that the 'manipulator' and the 'instrument' cannot be put on a par as agent in the same episode. In other words, "different ways of structuring a given situation cannot be activated simultaneously to produce a single coherent image" (Nishimura 1993:499). If one wants to expand the lower level episode to include the ultimate manipulator, a second constellation is needed, as has already been illustrated with Nishimura's example The poison which she had slipped into his coffee killed him. In this example, the ultimate agent, the person slipping the poison into the coffee, is secondary to the main episode, she is an actor in a by-play incidental to the main action (Nishimura 1993:496). Thus, the coding resolves the relationships between the Actors and the poison, or better, between their respective actions. When it comes to the lowest level action, the killing, it is still the poison which unequivocally is held responsible. Thus, I take instrument subjects as illustrated here not to be a mere surface phenomenon, but as reflecting a clear instance of recategorization, as also pointed out by Schlesinger: "the instrument by which a given action is performed may be conceived of as Instrument or, under certain circumstances, as a sort of Agent performing the action" (1989:205). As observed, these circumstances may involve a number of different, possibly interlocking factors, such as the lack of knowledge about entities further up the causation chain, the percep34 Ryder's construct of 'episode' is in fact equivalent to Langacker's 'scope of predication', DeLancey's (1991) 'event schema', and Nishimura's (1993) 'coherent action sequence'.


105

tibie salience of the instrument vis-à-vis the manipulator, the potential for independent action or other inherent properties of the instrument. These circumstances condition the feasibility of an entity to be conceptualized as Agent. Naturally, this view has a direct impact on the definition of the Agent category, since some entities will be better Agents than others and so no univocal, all-embracing definition can be given. The Agent category is a prototype-based category centring around a volitional Agent in control of his action, a "simultaneously ultimate and proximate cause of the event" (DeLancey 1984b: 184). Non-volitional Agents are extensions from the prototype. When acting involuntarily, the Agent is only a mediating cause as the act must have been instigated by something else. In the case of inanimate Agents, the issue of volition is irrelevant. Essentially, they come in two kinds, (1) those that are not known to have been manipulated by a more ultimate Agent, like for example accidents or 'Forces' and (2) 'Instrumental Agents' which are the immediate, but not ultimate causes. Note, however, that the distinction may not always be clear. As Nishimura observes, "once contextually furnished with force, any inanimate entity can be conceived to be on an equal footing as typical instances of Force" (Nishimura 1993:504). This is, once more, borne out by the data for Ml: the enormous variety in the Actor role makes it virtually impossible to provide a concise overview; in principle any entity or any event can become the Actor of kill We will return to this in our discussion of kill below. The prototype-based structure of the Agent is evidenced by language use. As is typical of prototype-based categories, not all members have the same degree of representativity; some instances of the category are more typical members than others. As already mentioned, this also holds for the Agent. Ryder points out that the linguistic tradition confirms that these differences are real, since it appears that "while some event participants will be considered agents by everyone, other, less prototypical ones will be considered agents to the extent that they can be construed as matching or approximating the prototype" (Ryder 1991b:300). Variations of judgment are thus a natural consequence of the nature of the notion of Agent. Against the background of the above characterization of agentivity, the remainder of this chapter will flesh out how the transitive paradigm takes the Agent as conceptually most basic building block. First, it will be shown how the MURDER verbs, a markedly transitive and prototypical subset of the field of killing, as it were incorporate a blue-print of Agentivity in their semantic structure. By bringing in the derivation of agentive nominals via the -er suffix, the chapter illustrates how the semantic network of this suffix parallels the prototype structure of the 'clausal' Agent. The lexical-paradigmatic analysis of this derivation, which at one point also draws another parallel with the

106


description of the middle construction in the preceding chapter, allows for a more coherent account of the semantic complexities of the -er suffix. The last section in this chapter considers in more detail the construction which could be regarded as reflecting the conceptual centre of the transitive paradigm, viz. the objectless transitive. Once again, a lexical-paradigmatic approach allows a more coherent explanation which does justice to the semantics of the construction and that of the instantiating verb. As such, it refines Davidse's analysis which is too grammatical as well as Rice's too lexically oriented description of this construction. 5.2 Lexical blueprints for agentivity: the MURDER verbs 5.2.1 Overview of the MURDER verbs Before considering in more detail how the Agent is instantiated by the each of the MURDER verbs, it is useful to give an overview of the selected items and their frequency in the database on which this analysis is based. Such an overview is given in Table 3. LEXICAL ITEM

kill murder massacre execute assassinate slaughter butcher slay lynch TOTAL

FREQ

PCT

794 245 104 78 77 31 16 14 4 1363

58.3% 18.0% 7.6% 5.7% ' 5.6% 2.3% 1.2% 1.0% 0.3% 100.0%

Table 3: Frequency of the MURDER verbs It should be noted that subsumed under the heading 'lexical item' are all verb forms, nouns (e.g. killing, murder) and compounds (e.g. lynch-mob). It can be reiterated that we do not aim at providing a full-fledged statistical analysis, yet we believe that the frequency of the items as well as that of their different meanings (not tabulated in Table 3) are important. As could be expected, the frequency of kill drastically exceeds that of the other verbs. Clearly, its broader semantic coverage, both in literal and metaphorical usage, explains the high frequency. The verb is also less tied to the transitive prototype involving a volitional Actor and a human Goal. Strikingly, when a frequency count is restricted to literal usages with a human Goal (in other words, 'proper' killings), it is the frequency of kill which drops substantially and the percentage of the other items increases. Kill also shows manifestly


107

greater variety in the types of Actors, with a high number of inanimate Actors, and often it is left unexpressed or is only vaguely defined. Often, the Actor remains so schematic that it seems inappropriate to speak of an Actor at all. It is in these cases that the distinction between Causes and Agents can become problematic. In contrast to kill, the other MURDER verbs incorporate into their semantic base an undeniable blueprint for an intentional Actor. This point is also raised by Dillon (1974:224) who says that certain verbs of "wrongdoing" require a "volitive NP". However, the semantic feature which entails this requirement is different for each verb, which further sets the items apart within the field. Moreover, different meanings of the verbs appear to influence the type of Actor selected. The following descriptions quite briefly illustrate the semantic structure of each of the relevant verbs, primarily (but not exclusively) focusing on the blueprint for Actorhood as incorporated into their semantics. The discussion of kill, the last one in the series, is considerably more elaborate for the obvious reason that it is more flexible in its lexical and constructional possibilities. By considering the more general verb kill, the discussion in fact widens the perspective to other causative verbs, e.g. punch or cut, for which the notions of Agents and Causes can be problematic. 5.2.2 Murder Undoubtedly, within the group of MURDER verbs, the strongest imposition of intentionality occurs with murder, which attributes to the agent a notion of premeditation, hence the verb's connotation of crime. It should be observed that murder imposes this premeditation regardless of whether or not this 'fits' the event. An example is someone charged with and convicted of second degree murder (causing the death of someone by neglect or refusal of help). Consider also the following example: (16)

... researchers ... accused of committing Jonestown-style murder on people with AIDS by not increasing the pace and scope of research and testing. (WSJ)

Clearly, the researchers do not have the intention of murdering people; the activist group puts it this way to wake researchers up to the urgency of finding a cure for AIDS. The specification of malicious afore-thought obviously determines the type of agent that can occur with murder: only entities capable of volition can be Actors to murder. This is unambiguously borne out by the data: in literal usage, overt agents are either human or human-like. The latter group comprises agents like company, government or army, entities that can be held liable for taking an

108


action. In these cases, there is a close metonymical between the collective and the people behind it. There are straightforward test frames which demonstrate the inherent intentionality οf murder: (17) (18)

*They murdered him, but didn't intend to. *They accidentally murdered him.

These constructions are clearly impossible, an occasional cynical use notwith standing. Even in the latter case, in which the notion of accidentality will be hedged either by intonation or explicit markers (e.g. quotes, as in They 'accidentally' murdered him), volitionality is asserted: they still imply that the murderers acted intentionally but do not admit to it.35 5.2.3 Assassinate A high degree of intentionality is also incorporated in the semantic structure of assassinate. A marked difference with murder, however, is the motivation behind the killing: whereas murder is prototypically characterized against the background of interpersonal relationships and individual motives (e.g. jealousy, greed, envy, etc.), assassinate finds its semantic base in the socio-political domain, implying political and/or ideological motives for the killing. Hence, the target of the killing is an individual of some importance within a certain social or political context (which explains why the item occurs relatively frequently in the WSJ corpus). The socio-political status of the patient finds its linguistic reflection in the fact that these patients are mostly identifiable individuals whose name is explicitly mentioned or for whom additional individuating information is added. Thus, assassinate requires the patient to be of some importance within a specific context, and for this reason the individual has been selected as the target of a hit, planned and controlled by an Actor. In fact, one could loosely define assassinate as "murder of an important person for ideological reasons". The type of Actor that can occur with assassinate is consequently limited to humans or human-like entities, acting intentionally, as witnessed by the unacceptability of *They assassinated John, but didn't intend to or *They accidentally assassinated him. Note that these sentences are acceptable when indicating that the assassins killed the wrong person; the unacceptability pertains to the denial of the ACTOR-PROCESS nucleus. This confirms the claim that intentionality is not a feature restricted to the Actor (or Agent) itself but applies

Recently, I came across the use of murderous accident in reference to the accident in which Lady Diana died. Conspiracy theories have it that it was a set-up.


109

to the relationship between the Agent and the Process (cf. Cruse 1973, Davidse 1991:29). 5.2.4 Execute In a sense, an ideological frame also lies at the basis of execute. The use of this verb in reference to a killing is an extension, or more precisely, a two step specialization, from another common meaning of the verb, viz. "to carry out, to perform". The first step in the specialization is the addition of the notion that the executioner is vested with the authority to carry out certain actions. This meaning lies at the basis of usages like to execute one's duties/power, executive director, etc.—usages quite common in the WSJ corpus. When the verb is used in reference to a killing, a second specialization is involved, this time applying to the type of action for which one is authorized, viz. the act of putting someone to death. Figure 15 provides a simplified representation of the two steps of semantic specialization:

Fig 15: A simple networkfor execute Since "to carry out a death sentence" is also an instantiation of "to carry out an action" a direct instantiation is also plausible. The direct instantiation loosens somewhat the notion that one is authorized to take the action, which reflects the semantic flexibility of execute as will be illustrated shortly. In the Western world, the power of execution is usually vested in the government as represented in the judiciary, operating on democratic principles. Hence, the killing is mostly interpreted as being justified and preceded by a fair trial. Yet language use may explicitly deny such a justification: (19)

... hundreds of men ... taken out at night in groups of five to 50 men to be executed without any judicial process whatsoever. (WSJ)

(20)

... the reported execution of U.S. Marine Lt. Col. William Higgins by his Shiite captors ... (WSJ)

110


Both sentences clearly report on executions ordered by some group in control, yet in both cases, without any trial, as example (19) explicitly acknowledges. Some people object to such usages of execute, like the reader of the WSJ who wrote the following in a letter to the editor: (21)

I take issue with the first sentence in your editorial about the 'execution' of Lt. Col. Higgins. ... The use of this term absolutely infuriates me. I thought perhaps I was wrong, so I checked my Webster's with the following results: Execute — "To put to death in compliance with a legal sentence Execution — "A putting to death as a legal penalty ..." Clearly, Lt. Col. Higgins was murdered, not executed. (WSJ)

The writer's reaction indicates that the meaning "to put to death in compliance with a legal sentence" is prototypical, especially in a Western conception of the world. The usages to which he reacts are extensions reflecting deviations on the basis of one's authority to execute. The executioners (not accidentally calling itself 'Revolutionary Justice Organization') assume they are authorized to kill Higgins, with which not everybody will agree. What counts as legally and morally justified is amenable to reinterpretation, it is essentially a matter of judgement and ideology. Now, interestingly, the type of Actor tends to vary in correlation with these two different usages of execute. In usages where it is used in reference to the sentencing by a tribunal, the agent is mostly a depersonalized institution like a state or a government. On the other hand, when the execution finds less moral or legal justification, there is a tendency to use more individualized agents, shifting the responsibility to these individuals. Further, the corpus material reveals that in reference to death sentences with little or no legal or moral justification (from a Western point of view), one more frequently has lexical variation, whereas this is absent in the case of 'legal' executions (ignoring some neutral rephrasings as to put to death). This follows logically from the interspeaker variability of what is deemed legally and morally justified. Mostly, the lexical variation more explicitly expresses the absence of justification. A case in point is the aforementioned killing of Col. Higgins, an event frequently referred to in the WSJ.36 Many references to this event (some 20%) are cast in terms of murder, as in example (21).

Such events have been tagged with a unique identifier, so that references to these can be singled out for comparison, as is done here.


111

In short, it cannot be denied that for all uses of execute, the Actor (institutional or individualized) is saliently volitional: the Actor consciously decides, within the lines of a given ideology, to put someone to death. 5.2.5 Lynch Maximally opposed to the prototypical meaning of execute is the meaning of the low-frequency item lynch, referring to a killing without due process of law. It is related to the extended use of execute as in (20), yet with lynch the implication is that more than one Actor is involved, often pictured as a raging mass of uncontrollable individuals lusting after revenge. This, too, implies a form of volition albeit one that is perhaps more subconscious. Hence, although marked, it is not impossible to say They lynched him, but didn't intend to, since the lynch-mob may only come to their senses after the event. Given that the item has a low frequency and a quite specialized meaning, little can be said about it, except perhaps a tendency for it to occur in metaphorical usage. 5.2.6 Slay Slay is often said to be a stylistic variant of other more common expressions of killing, usually reserved for a more formal register. If this is true, then it is to be expected, first, that it will predominantly occur in variation with other expressions of killing in reference to the same event (e.g. within the same text). Secondly, the expectation is that slay will be rather infrequent, since it appears chiefly in contexts where stylistic variation is called for. Both expectations are generally confirmed by the data. Once more, the WSJ corpus provides interesting material to test these hypotheses, since it contains frequent references to the same event (like the Higgins killing). Table 4 gives an overview of the lexical variation that has been attested in the WSJ for such identifiable events that are also coded by slay. LEXICAL ITEM

kill shoot assassinate massacre die execute murder TOTAL

FREQ

PCT

8 4 2 1 1 1 1 18

44.4% 22.2% 11.1% 5.6% 5.6% 5.6% 5.6% 100.0%

Table 4: Lexical variation for slay

112


Usually, slay is the second coding in the discourse, which even more strongly suggests that slay is primarily a stylistic variant and is not a prominent coding in and by itself. Slay is a markedly transitive verb which calls for an intentional Actor, at least in contemporary English. Presumably, the strong transitive character and the common implication of deliberate cruelty associated with it are features inherited from the item's Old English ancestor sléan which means "to smite, strike, or beat". Quite early, the word acquired the meaning of "to strike or smite so as to kill", a natural evolution, especially in view of the fact that in those days beating and/or hitting weapons were most common. In contemporary English, slay is no longer a verb of forceful contact (with the strong implication that the contacted entity is killed) but a verb of killing (with arguably a strong implication that the entity killed is forcefully hit). The shift from "beating" to "killing" is a natural one and by no means restricted to slay. Consider, for instance, the use of hitman in reference to a (hired) assassin or the use of hit in reference to a murder. Interestingly, the OED data show that slay was quite flexible: from the 11th century to the late 19th century the verb came to express deaths resulting from accidents, natural disasters, etc. or 'official' executions. Only the latter is still possible in contemporary English: (22)

London has concluded that Austrian President Waldheim wasn't responsible for the execution of six British commandos in World War II, although he probably was aware of the slayings. (WSJ)

Strikingly, also metaphorical extensions, such as "to bring to spiritual death", "to destroy (the vitality of)", or "to suppress completely" were possible with slay. In other words, the OED data indicate that slay manifested the same kind of semantic variability as present-day kill does (cf. below), yet for some reason it lost this flexibility around the end of the 19th century. In fact, for several hundred years, there must have been some overlap between slay and kill. A more elaborate diachronic analysis is necessary to reveal in which respects the items still differed and what caused the eventual semantic specialization of slay. Such an analysis is beyond the scope of the present discussion.37 At this point, we can take it that the contemporary use of slay has retained what had rapidly evolved to the item's prototypical meaning, viz. the deliberate killing of a human being. The notion of intentionality (as well as the often One thing I noted in the OED data is that slay appears more constrained than kill (both diachronically and synchronically) in the type of affected entity that can occur with it; it is usually a human being, even in metaphorical usages. This may also be inherited from the item's original prototype of direct contact with the Affected.


113

implied cruelty) may be attributable to the original centrality of beating and/or stabbing weapons which are willfully controlled by the assaulter. 5.2.7 Massacre In the set of MURDER verbs, massacre stands somewhat apart, as it indicates that many people are killed. The plurality of the Goal and often also of the Actor is a prerequisite, since constructions like *The army massacred {a/the} student or *He massacred her are unacceptable. The salience of plurality appears to motivate the high number of nominalizations as well, all of which evoke a holistic conception of the event and profile the result. Verbs, in contrast, designate sequential scanning, which entails a more individuating conception of the people killed. In the case of a coding with massacre, such an individuating conception may be more difficult. The difficulty of individuation also explains why the agent, when overtly expressed, is nearly always coded by means of either a collective like army, government, country (in total approximately 70% of the cases) or a plural noun. If the Actors of massacre are predominantly soldiers and the like, then the question arises whether they are still intentional Actors. Our knowledge of the real world tells us that soldiers carry out orders given by their superiors, who are thus in fact the ultimate malevolent causers of the event. In a coding with massacre, the grammar does not reflect that distinction, except when this ultimate causer is explicitly coded as the Actor. When the soldiers are the Actors, the superior officer is outside the scope of predication and the Actors are seen as solely responsible for the event. The Actors associated with massacre are conceived as having intentionally initiated the act and as controlling it: (23) (24)

*They massacred students, but didn't intend to. *They accidentally massacred students.

5.2.8 Butcher and slaughter Like massacre, both butcher and slaughter may denote a large-scale killing of human beings, yet unlike massacre, they do so by extension. In essence, both items emphasize the atrocity of the killing, which is readily accounted for by their link with the killing of animals. The extension to a large-scale killing of human beings is motivated by the fact that such an indiscriminate killing shows little compassion for the victims, as if they were no more than animals. Unlike massacre, the plurality of the (human) patient is not an absolute prerequisite: (25) (26)

It's just that—you're going to slaughter me ... (LDC) I couldn't help seeing ... that [young boy] lying butchered (NOV)

114


However, such usages are markedly less frequent, and reference to massive killing is very common, more so for butcher than for slaughter. The data from the OED indirectly confirm this tendency: in reference to the killing of humans, butcher is never attested with singular patients (although this is a possible coding); for slaughter in contrast various attestations with singular patients are listed. Further, slaughter is the less explicit of the two and, therefore, also more frequent in its reference to the killing of humans. Is the difference between the two items a matter of pragmatics or mere convention? I have reasons to believe that the difference arises from the distinct semantic structure of the items themselves. These distinctions also influence the type of Actor that may figure in the conception of the process. The etymologies of slaughter and butcher go a long way towards indicating where their semantics differ. Slaughter is etymologically related to slay, and inherits a notion of forceful contact with the patient. In present-day usage, this is perhaps no more than a hint. Butcher, in contrast, is a verbalization going back to Old French bouc, "he-goat", from which boucher "vendor of goatmeat" is derived (Partridge 1979; cf. also Clark and Clark (1979) who mention butcher as an example of an 'agent verb'). So, in Middle English both the concept 'butcher' and 'meat' figure prominently in the conceptual cluster associated with butcher, and are still salient in the present-day complex. For instance, the place where animals are killed is typically called a slaughterhouse, but the place where you buy meat is the butcher's. So, the butcher prototype can be described as "the killing of animals in the manner of a butcher, cutting them up, dressing them, and selling their meat" whereas slaughter's can be paraphrased as "killing animals, esp. for food". When extended to the killing of human beings, then, the difference is carried over; butcher more prominently evokes the image of wounded bodies and blood. The following use of butcher (in reference to abortion) juxtaposed with bleed is suggestive: (27)

DOCTOR: Will you, ahem ... VIRGINIA: That's my business. I'm not a kid. Old enough to bleed, old enough to butcher. (LDC)

Given the few attestations for slaughter and butcher, generalizations are somewhat unreliable, although the metaphorical uses seem to align with the distinction: when used metaphorically slaughter tends to emphasize the large scale effect (e.g. in the WSJ, stock traders are said to have been slaughtered in a crash), whereas butcher tends to emphasize the destruction (e.g. a text can be said to be butchered by an actor's bad delivery). Yet even here caution is warranted. Also, the difference seems to finds its reflection in the type of Actor: although both verbs typically select a human volitional Actor, butcher is much


115

more constrained to this type of Actor than slaughter, as evidenced by the ambiguity in the following attestation: (28)

[Hezbolla] planned the bombing of the Marine barracks near Beirut in 1983, slaughtering 241 Marines (WSJ)

The ambiguity is due to the non-finite participial construction: the subject to slaughtering can be either the subject of the main clause {Hezbollah) or the object {the bombing). Further illustration and discussion of this type of ambiguity in non-finite clauses will be offered in the next section, as it is quite common with kill. 5.2.9 Kill Ironically, the verb which occurs most in the literature as typically transitive, the general verb kill, conforms least to the intentional Actor prototype of transitives. Its generality as well as non-prototypicality are explained by the interaction of a number of factors: (1) its high lexical flexibility, (2) a less stringent implication of goal-achievement, (3) the possibility of a non-volitional Actor and (4) the possibility of an inanimate Actor. The following part takes up each of these factors in turn, with most emphasis being given to factors (3) and (4) as they pertain directly to a proper account of agentivity which is argued to be the backbone of the transitive paradigm. The decription of the types of inanimate Agents possible with kill further refines and substantiates the introductory description of the Agent as a prototype category. It also takes the discussion to a more general level, considering the relationship between Agents and Causes. Our view on this relationship follow Davidse's characterization and departs from the cognitive view. Lexical Flexibility. A first dimension in which kill is markedly different from the other MURDER verbs is its considerable lexical flexibility, which supports the intuitive categorization of it as a hyperonymical item, i.e. one that can be used to refer to the entire lexical field. This is directly borne out by the corpus material: the portion of metaphorical uses of kill averages around 22%, whereas for most other MURDER verbs metaphors generally amount to less than 10%. Moreover, some metaphorical uses of kill have become quite conventional. Here are a few typical usages, each time illustrated by attestations from the corpora: (a) (b)

to halt a process e.g. to kill the production, to kill the peace process to stop the existence of laws, regulations, etc. in effect e.g. to kill the catastrophic surtax

116


(c) (d) (e) (f) (g)

to prevent an entity from materializing/taking effect e.g. kill a proposal, kill a bill, it killed the speech (not yet delivered) to stop the existence of abstract objects e.g. this kills the American dream, to Ml the human spirit to incapacitate (hyperbolical use) e.g. my feet are killing me to take drastic action (hyperbolical use) e.g. housewives who would Ml for her job getting rid of things in excess e.g. kill the details, Ml some time

It would take us too far to consider all the possible extensions of kill and their mutual relationships. Against the background of the earlier description of the lexical extensions of kill (Chapter 3, section 3.1), it will not be difficult to discern some major semantic links, which, however, I leave as an exercise to the reader. Goal achievement. In addition to its high lexical flexibility, Ml is also nonprototypical (and thus different from the other MURDER verbs) in being less tied to goal-achievement. Nishimura (1993:506) observes that the meaning of a verb denoting goal-directed action can generally be decomposed into an action component and a goal-achievement component. In the prototypical case, the two components are fully activated, yet in other cases, one of them may be suppressed. With inanimate Actors, for instance, the focus shifts from the action component to the goal-achievement component. In fact, as Nishimura points out, "the achievement of the goal is almost invariably a necessary condition for an inanimate or abstract entity to be viewed as Agent" (1993:507). In other less typical instances, the goal-achievement may be suspended. A device often exploited for coding such construals is the progressive aspect. Davidse (1991:34) cites the following example: (29)

Help! That dog is biting me!

In the usual interpretation, it is assumed that the dog actually has its teeth into you. Yet, you could well utter this sentence when the dog has not bitten you yet but when you believe the animal has the clear intention of doing so. In hyperbolical usages of Ml, as in my feet are killing me, the use of the progressive is similarly quite common. The absence of goal-achievement is essential, since your feet are not actually killing you but merely hurting very badly. The reason why the progressive implies a lack of goal-achievement is because it furnishes an internal perspective on the event, excluding the end


117

points from the immediate scope of predication (cf. Langacker 1991b:207ff). For instance, if one witnesses a violent assault one may say (30)

Call the police! He is killing him!

Interestingly, the progressive use seems to be restricted to kill and does not occur with the other MURDER verbs: (31)

Call the police! ?He is {murdering/assassinating/executing} him!

unless perhaps when one suddenly realizes that the assaulter had planned it all along.38 In other words, these verbs foreground the action as well as the goalachievement, which again indicates that they are more typically transitive than kill. The reason for the stronger focus on the goal-achievement may be that they strongly imply intentionality and/or premeditation, which also entails a focus on the goal as premeditated by the agent. If you plan something meticulously, then what you plan becomes as real as the actual event.39 My claims are also supported by the type of nominalization that has become conventional for the different MURDER verbs. Without getting too much involved in the difficulties of nominalization, it can be observed that zeronominalizations (e.g. murder) and nominalizations in -ion (e.g. execution) differ from -ing nominalizations precisely in the salient evocation of the result of the process. The former profile the event as a whole, often shifting to designating the result (e.g. a crack, a hit, a homerun, a touchdown), whereas the -ing nominals furnish an internal perspective on the process, excluding the end points (thus revealing the semantic parallel with the progressive, cf. Langacker 1991b:Ch.l). Strikingly, kill is the only MURDER verb for which the -ing nominalization, which furnishes an internal perspective, is the most conventional one. All the other MURDER verbs take either a Ø-nominal or a full episodic noun in -ion (e.g. assassination, execution, etc.). The different tendency in the nominalization further supports the view that goal-achievement is much less prominent for kill than for the other verbs. Logically, when the 0-nominalization applies to kill, it also strongly implies goal-achievement, which is what gives the (less 38

The progressive does occur with other MURDER verbs, e.g. Let's run away! They are {executing/assassinating} Western visitors (Brygida Rudzka-Ostyn, pc). However, in these construals, the process is being situated on a large scale (as part of a policy being carried out) and iterative aspect is involved here. As such, they are different from the examples cited here, which involves the internal perspective on a single event. 39 It can be noted in passing that the present view is diametrically opposed to that held by Kac (1972a,b) who identifies murder as "an action predicate" which means that "it asserts an action and implies a result". Kac characterizes kill as a "result predicate" asserting a result and presupposing an action "in at least some instances" (1972a: 123).

118


common) noun kill its connotation of greater cruelty. The following example is indicative: (32)

...Gilbert sets upon a particularly hapless rabbit and Mrs. Ford, as the sport demands, steps in for the final kill by stomping on its head. (WSJ)

The conception of the process as being bounded enhances the achievement of the goal, which in this sentence is further emphasized by final. Non-intentional goal-directed action. Assymetrical to goal-achievement is the notion of intentional goal-directed action. Unlike the other MURDER verbs, kill does not stringently imply any volition on the part of a human Actor. Consider the following examples: (33) (34)

... [Mr. Parks] isn't to blame because he killed the woman without "will or conscious mind." (WSJ) ... the House deliberately killed a handful of projects... (WSJ)

Mr. Parks killed the woman while sleepwalking and cannot have acted voluntarily, so the WSJ article reasons. The use of the adverb deliberately in example (34) confirms the item's neutrality concerning volition. This neutrality sets kill apart from the other MURDER verbs, which normally would not be used with deliberately'. (35) ?They deliberately {murdered/assassinated/executed} him. For reasons of emphasis, people may still decide to use such a construction, yet it is quite marked, as witnessed by the total absence of such constructions, or similar ones, in any of the corpora consulted. It can further be noted that only some 20% of the attestations with kill express verifiably volitional acts. I should note that in contrast to the transitive model, the ergative paradigm has no central hook-up with intentionality, as observed by Davidse (1991). In the corpus material for prototypically ergative verbs of killing (e.g. suffocate, drown, or starve), a human Instigator occurs in barely 30% of the ergative constructions, a marked contrast with the MURDER verbs for which the ratio of intentional Actors reaches up to 95% when kill is excluded and approximately 60%) when it is included. Davidse (1991) points out that even with a human Instigator the accidental interpretation is as natural as the deliberate one: (36) (37) (38)

I'll try not to drop bricks. I broke that vase cleaning. You're spilling the water.


119

While this indeed seems to be the case, I nevertheless extrapolate from my corpus that when an overt human Instigator surfaces with ergative verbs of killing, intentionality is almost invariably implied. If the Instigator acts involuntarily or is more external to the event, there is a strong tendency to use an non-effective construction, possibly within an analytical causative, as in the following example: (39)

The first liberal reaction to the famine was to blame President Reagan for "letting Ethiopian kids starve because the regime was communist." (WSJ)

Notice that in contradistinction to the effective construction (IS-PROC-ME), the let-construction in (39)40 and make/get-causatives retain the voice vagueness of the non-effective. I return to the characteristics of the ergative effective construction, as well as the conditions that influence its occurrence in Chapter 6, when discussing the predominantly ergative SUFFOCATE verbs. Inanimate Actor. Although with kill human and human-like Actors are still prominent (approx. 55%), the fact that the verb does not stringently presuppose a volitional Actor has paved the way for inanimate Actors, which indeed occur quite frequently. When inanimate Agents occur with causative verbs, linguists invariably bring in the notion of Causes, as opposed to Agents. In the next part, we first describe some of the features of inanimate Actors with kill after which we will evaluate the relationship between Agents and Causes more critically. The type of inanimate Actor that occurs with kill is spread more or less evenly between 'instrumental' Actors (48%) and events (52%). As was pointed out in the introductory discussion of Agents, language users must decide at which level they want to express a given event, and must choose which entity they want to code as the Agent directly responsible for the event. It was further noted that this choice is an either-or affair, in the sense that 'manipulators' and 'instruments' cannot both be coded as Agent in one and the same construal; recall the unacceptability of example (15), *The antibody and scientists killed AIDS-infected cells. However, the data for kill indicate that things may not always be so clear-cut. Consider the following alternate constructions with kill, all related to abortion:

40

The let-construction, implying permission rather than coercion, is perhaps more accurately characterized as a permissive rather that as a causative (cf. Verhagen & Kemmer 1994).

120


(40)a. ... because it [RU-486] kills the unborn ... (WSJ) b. ... [RU-486] is now being marketed to kill unborn babies ... (WSJ)

c. ... had anyone killed me, in or out of the womb ... (WSJ) While (40a) explicitly codes the abortion drug as agent, (40c) casts the ultimate human causer (anyone) into the agent role. As such, they represent different episodes of similar events.41 Example (40b) straddles the border, since one can interpret the implied agent of infinitival kill to be either the drug (mentioned in the main clause) or the humans using the drug. Regarding such type of coding, Ryder states that "there are human agents indirectly causing the reaction of the patients, but they are all distant from the event" (1991b:307). Ryder's observation is essentially correct, yet this still leaves us with the question of who or what is the Actor in (40b). It appears that in (40b) or similar constructions, neither the drug nor its implied user can be identified as the 'true' Actor, although the contextual focus on the drug makes it a more prominent candidate. The reason for the ambiguity is the non-finite construction, as was also noted to occur with slaughter. By not being tied to a finite verb, either the drug or the humans using it can be interpreted as Actor. Similar ambiguities occur when an event is coded as the Agent of the killing. If an action or event is construed as Agent, the causative relation shifts from a relation between a single participant and a single action to a relationship between two events. Such a conception of causation lies more in line with a philosopher's point of view, in which a causative relation is said to link two events, where one event is seen as bringing about another event (cf. for instance Davidson 1980). 'Event Actors' figure prominently in the expression of accidental killings, as in the following examples: (41) (42)

Flight 173 ran out of fuel and sputtered into a forest, killing 10 people, including two crew members. (WSJ) Later, Gur was killed by a playmate's accidental discharge of an antique shotgun. (WSJ)

Example (42) illustrates how the playmate's role in the event is downgraded due to the fact that his action was accidental. Ambiguities in the identification of the Actor mostly occur in two types of constructions. The first is the serial verb construction, like This is not entirely correct, given that (40c) applies to other types of abortions (i.e. not induced by RU-486), but for the sake of the argument this can be ignored. Note further that the speaker casts herself into the role of the unborn, which also brings in an element of contrast, but this, too, I choose to ignore at this point (see, however, Chapter 7, section 7.4)


(43) (44)

121

Israeli troops shot and killed an Arab activist (WSJ) Interstate 880 ... collapsed and killed more than 200 commuters (WSJ)

In these constructions, it is not the Actor tout court but rather the Actor as engaged in a specific process that effects the death of the Goal.42 The same is true for constructions like dozens have been killed by exploding mines, where the focus is on the mines' independent activity. A second case where ambiguities arise due to co-activation of Actor and event is, once again, a non-finite construction, as in the following example: (45)

The group has twice bombed U.S. embassy buildings in Beirut, killing a total of 72 people ... (WSJ)

Also here, either the event denoted by the main clause or the group participating in it can in fact be regarded as the Actor of killing.43 Clearly, participial constructions need not always be ambiguous, as already illustrated by (41) which unequivocally takes the crash described in the main clause as the Actor. In other words, in these cases there appears to be some competition between the human Actor (who in the typical event conception is more salient than his action or the instrument he is using) and the Actor's action. Note that this is tied to the fact that kill commonly occurs with 'Event Actors'; verbs which saliently incorporate the notion of a volitional Actor, like the other MURDER verbs, are not ambiguous, as the following example shows: (46)

The group has twice bombed U.S. embassy buildings in Beirut, {murdering/assassinating} a total of 72 people ...

Codings with an 'Event Actor' are, however, most common when the event cannot be traced to an initial Actor, as is the case with accidents, general acts of violence (such as warfare or bombings) and diseases. Especially accidents frequently occur in the Actor role with kill: in almost 45% of reports on deaths in accidents, the accident is coded as the Actor; conversely, more than 60% of the 'Event Actors' are accidents. Another common coding for accident casualties is an agentless passive construction; it occurs in almost 50% of reports on casualties. In these constructions the Actor remains indeterminate. The WSJ corpus, once again, provides an interesting source for comparison, as 42

See for instance Shopen (1977), Lakoff (1986) and Jaeggli & Hyams (1993) for insightful ideas about the co-ordination construction. 43 This is a type of 'notional subject' that has often escaped the attention of grammarians. For instance, Quirk et al. (1985) and Declerck (1991) focus solely on constructions with a notional subject that is also a constituent in the matrix clause.

122


there are multiple references to identifiable accidents. As is typical of newspapers, it often adds the number of casualties whenever the event is mentioned. Let us consider one example, a set of references to the crash of a jumbo jet forced to land in a cornfield in Sioux City, Iowa: . (47)a. ...what caused the crash that killed 112 people. (WSJ) b. [the failure] that caused the United Airlines jumbo jet to crash last week in Iowa, killing 111 people. (WSJ) c. One hundred eleven people were killed. (WSJ) d. In that disaster, in which 111 were killed ... (WSJ) The (a) and (b) sentences construe the event as Actor, the (c) and (d) sentences illustrate agentless construals. Notice that while for (c) the crash could still be interpreted as the omitted agent (111 people were killed by the crash), this is not possible for (d): *in that disaster, in which 111 people were killed by the crash. This is quite normal given that within one episode something cannot simultaneously function as setting and as agent. The question, then, is what is to be regarded as the Agent in the 'circumstantial passive'-construction in (d)?44 None of the corpora (including all the 427 passive constructions in the OED) contains any occurrences of an overt Actor with such a 'circumstantial passive'construction. The absence of overt agents in these constructions is experientially grounded: in events like the ones reported on, there is usually no agent more salient than the event itself. If anything is to function as agent in a circumstantial passive construction, then it has to be a participant situated within the event and one which furthermore acquires sufficient prominence to be construed as agent. In addition to one corpus attestation that comes close to such coding, a more explicit construal combining an overt Actor with a circumstantial passive is sentence (49): (48) (49)

The incubator blew up on them, and they were ... killed by the shrapnel (CASS) In the explosion, 7 people were killed by flying shards, 1 person died from a heart attack.

Such usages specifically identify what exactly caused the deaths of the people that died in the accident. While possible, such overt shifts from the setting to an Actor within that setting are infrequent, as indicated by the data. The following 44

The term 'circumstantial passive'-construction is a convenient cover term for the type of construction, as in (d), where an agentless passive is used and the event surfaces in a periphrastic coding, usually as a setting. Included, however, are also other types of circumstantial codings of the cause, e.g. The others were killed ... when the plane demolished their homes. (WSJ).


123

simplified diagram visualizes the different constructions. For contrastive purposes, I include a representation of a related, yet different construction, to die in a crash. (a)

(b)

die in a crash

(c)

be killed in a crash

be killed by a crash

(d)

the crash killed 112 people

Fig. 16: Settings and Agents In all four constructions, the event is nominalized and consequently represented as a thing (indicated by a circle). The difference between (c) and (d) represents the semantic purport of the passive construction, viz. a trajector/landmark reversal (Emanatian 1993; cf. also Langacker 1987a:120ff). The possibility of (d), as in (50)

... the crew were killed by the same blast ... (CASS)

presents counterevidence to Dirven's (1993:95) claim that ''killed by [...] is incompatible with circumstantial causes such as accidents" and that consequently the circumstance can be expressed only by means of in, as in be killed in an accident. When accidents or the like are coded as an Actor, as in (47a) or (50), they are conceptualized as directly participating in the process and affecting the Goal. It remains a fact, however, that this type of construction is less typical... Usually a circumstantial passive construction will be used, represented by (d). In this construction the landmark remains schematic, as indicated by cross-hatching. In some contexts, the landmark may be elaborated (cf. example (49) above), but usually it is not. Regarding event structure, the schematic landmark is what distinguishes (b) be killed in a crash from (a) to die in a crash. 5.2.10 Causes and Agents The relationship between kill and die in the context of accidents or diseases can be puzzling. DeLancey (1984b), for instance, who takes agentivity to fundamentally involve the causation of an event, concedes that he does not see a clear semantic difference between the constructions in a pair like (51), to which, in view of the preceding discussion, I add (52):

124


(51)a. b. (52)a. b.

The woman died from a cold. (DeLancey 1984b) A cold killed the woman. a second worker died from an explosion and a fire ... (WSJ) ... the Challenger explosion, which killed seven astronauts ... (WSJ)

Nishimura (1993) dispels DeLancey's doubts by pointing out that in (a) cold—and consequently also explosion in (52a)—are Causes, whereas in (b) they are Agents. Inanimate Agents, Nishimura says, can be subject to recategorization as Cause and vice versa. This leads him to align himself with the tradition of regarding the prototypical Agent as constituting the most salient instance of Cause.45 It is precisely with this view, which also in Cognitive Grammar has become widely accepted, that I intend to take issue, aligning myself more with Davidse's views on the issue. Characterizing the prototypical Agent as a prototypical Cause not only fails to explain the semantic difference between the two, but also makes incorrect predictions. If this view is correct, it should in principle be possible to use them interchangeably, which, however, is not the case: (53)a. *The sniper died a worker. b. *One worker died {from/of} a sniper. c. *One worker was killed {from/of} a sniper. The unacceptability of these constructions reveals a semantic difference between Agents and Causes, which in English finds an additional reflection in the use of different verbs. In the case of die, there is only one participant, the intransitive Actor; the cause is expressed periphrastically by a from or of-adjunct and has no participant status whatsoever.46 The semantics of the prepositions from and of is clearly relevant here. In constructions like to die from/of something, they extend from their locative prototype and construe the causing event metaphorically as the abstract temporal origin of the causative chain. However, when the same ' causing event surfaces in a coding with kill, it cannot be conceptualized in these 45

For example, Lyons (1977:490) claims that "agents are seen as the causes of the situations which, by their actions, they bring into existence". He further emphasizes the "natural connexion" between causality and agency, but the former is still "logically distinguishable from agency". Unfortunately, he never explains what this logical distinction is. More recently, Givón (1994:7), following Hopper & Thompson (1980), identifies the agent (of a prototypical transitive event) as "salient cause" in polar opposition to the patient, which is "salient effect". 46 Following Talmy (1985b), thefrom/of-adjunctis taken here as expressing the cause (cf. also DeLancey 1984b:204ff). Clearly, some constructions are possible with both verbs, e.g. Many people {were killed/died} in the accident, but here the in-adjunct expresses a setting, which can be added to practically any action or event.


125

terms and is therefore promoted to an active participant that directly effects the result. As such, kill (like any other transitive verb) is associated with an 'active' cause; even 'inactive' causes as 'm pneumonia killed her are rendered active by the agentive construal. In these cases, the meaning of the transitive effective construction "shifts from 'controlled targeting onto a Goal' to something like 'effective contact with or impact on a Goal' " (Davidse 1991:40) and the construction necessarily implies goal-achievement (cf. also below). The unacceptability of the sentences in (53) indicates that volitional Actors cannot be recategorized as Causes. As observed by DeLancey (1984b), they have to be coded as an agent to a transitive (=effective) construction. The notion of Cause only comes in with unprototypical inanimate Actors. So, following Davidse (1991), I restrict the notion of agency to direct participation in the process: the transitive agent function is instantiated by the Actor, the ergative agent by the Instigator. Causes have a different value, straddling the border between a nuclear participant and a circumstantial setting. This point of view finds a confirmation in Talmy (1985a) who distinguishes different types of causation lexicalized in verbs by contrasting patterns like those in (52) and (53). Relevant in the present context is that he attempts to distinguish "agentive causation" (possible with kill or murder but not with die) from "resulting-event causation" (possible with die but not with kill or murder). His terminology agrees with the position defended here. It has already been observed by a number of scholars47 that imposing an agentive image on the cause—as entailed by the effective constructions in (51b) and (52b)—personifies the cause to some extent. Some clear examples of such personification are the following: (54)

(55)

It started when Frank's CD player tried to kill me on my way to work. ... [ML: Frank throws his CD player out of the window] ... "Frank!" I yelled up at the open third-floor window. "You could've killed me with your goddamned CD player!" (CASS) Failing to drown me, the wall of water decided it would just try to bash me senseless against the street. (CASS)

In these contexts, objects are personified to the extent of becoming like volitional Actors (note the use of tried and decided). Being typical of creative prose, such examples stand out much more than those attested for example in the more formal prose of the WSJ, where an 'agentive cause' in the role of Actor is quite frequent and quite unmarked. Significantly, for instance, in relation to reports on casualties in accidents discussed earlier, the WSJ data reveal that kill

Cf. for instance DeLancey 1984b:197; Nishimura 1993:504-5; Davidse 1991:77.

126


is used much more frequently (72% of the cases) than die (23%). For diseases, the ratio is reversed: kill occurs in 14% of the cases and die in 86%. Since the verbs overlap in the coding of both accidental deaths and those resulting from diseases, a watertight generalization is impossible to make. Nevertheless, the results presented here definitely support the generalization that kill will more readily be used for events that are prototypically caused by external and perceptually distinguishable causes, whereas die will be more typical in cases of less perceptible causation, often of the kind that comes 'from within'. I hold the view that in the prototypical case a transitive construal entails a maximally external point of view: the Actor is an entity external to the Goal and directly impinging on it, from the outside. Chapter 6 will substantiate that ergatives, in contrast, by definition imply a more internal perspective, the Medium being a secondary energy source. In an intransitive conception, as for instance designated by die, the participant is the sole energy source of the action and the cause is only remotely involved as the abstract location (the origin of the chain of events). The semantic structure of die specifies that the action is accidental, which relates the Actor to the non-volitional Actor of a transitive clause. As Davidse (1991:111) notes, the distinction between volitional and non-volitional action tends to be lexicalized with intransitives (o.g. jump vs. fall) which sets them apart from the transitives for which this distinction is generally not hard-wired into their semantic structure. Similarly, DeLancey (1984b: 181) says that "active intransitive verbs (run, jump, etc.) require of their single argument the same volition as agentive transitive verbs (throw, build) do of their agents". Since die codes an accidental process, the agentive nature of the Actor may be less than with volitional intransitive Actor, but nonetheless, the dying participant is an active participant, whereas in a coding with kill it is an inert Goal. Returning to the coding of accident victims, it should be added that in the case of the passive kill-construction with the accident coded as setting, the difference with die is minimal, and both codings often alternate ({be killed/die} in an accident). As argued above, in such constructions the 'agentive' aspect of the cause is quite absent. Significantly, if only these passive kill-constructions are retained in the comparison with die, the percentage of reports on accidents vis-à-vis the total of literal usages of kill is reduced to 10.5%), which is the same result as for die (10%), although in general a coding in terms of kill is still more frequent. An additional pragmatic factor for this, at least for the WSJ, is the unspoken law of media language that a report gains in newsvalue when people are killed instead of just having them die. From the preceding discussion of the semantic structure of the MURDER verbs, it is to be retained that all these items incorporate a salient blueprint for an


127

intentional Actor, granting some leeway for kill which is less committed to this type of Actor. This strict commitment to an intentional Actor makes these verbs more static, especially when seen in contrast to the enormous lexical and constructional flexibility that characterizes the SUFFOCATE verbs, particularly the prototypically ergative ones. For example, the OED does not list any ergativization for any of the MURDER verbs, a process which nevertheless appears to have been a productive one, affecting even the transitive SUFFOCATE verbs. The non-occurrence of the ergative pattern is undoubtedly due to the salient focus on the Agent that is 'hardwired' into the semantic structure of the MURDER verbs, making them prime representatives of the transitive paradigm. It has further been illustrated that kill is the most general verb of the field, since it can be used in contexts where other verbs of killing cannot. More specifically, the verb was shown to be more flexible in its lexical coverage, to be less committed to either goal-achievement or intentionality. As a logical consequence of the latter, the data revealed a high frequency of inanimate Actors. Moreover, they often appear in construction types that help reduce the markedness of inanimate killers. The above description is a first assessment of the Agent-centredness of the transitive paradigm as incorporated into the lexical structure of prototypically transitive verbs. The following two sections in this chapter widen the perspective to a more constructional one. First, the derivation of agentive nominals via the -er suffix will be considered. The Agent-category designated by this suffix parellels the one observed at clause level, but further paradigmatic complexities (e.g. the middle construction) should be considered as well. Secondly, the objectless transitive construction will be discussed and will be shown to maximize the focus on the Actor, which justifies the claim that this construction singles out what is conceptually basic to the transitive paradigm. 5.3 Morphological blueprints f or agentivity: agentive nominals Additional evidence of the structuring importance of the transitive and ergative paradigm comes from English morphology, viz. the derivation of agentive nominals through the -er suffix. Data from my corpora overwhelmingly indicate that the morphological derivation is indeed governed by the two paradigms. First, it appears that the -er derivation manifests itself primarily within the transitive paradigm, which is intuitively plausible given the Agentcentredness of this paradigm. The -er formation can thus be taken as confirming the Agent-centred character of the transitive model. Secondly, the -er suffix has traditionally been characterized as the agentive suffix, i.e. deriving a noun which refers to the Agent of the process designated by the base verb. In view of our characterization of the notion of Agent as a prototype category, this is essentially correct. However, the -er nominal may also come to profile the Goal

128


of a transitive process as well as the Medium of an ergative construal. Such formations can find a coherent explanation only if interpreted against the background of the transitive and ergative paradigms that shape the construal potential of the base verbs. Although my discussion of the -er suffix has benefited greatly from Levin & Rappaport (1988) and Ryder (1991a,b), I believe that the present account is an original contribution to the study of agentive noun derivation. It manages to explain frequency deviations overlooked by other researchers as well as to account for 'deviant' formations noted by others yet largely left unexplained. The discussion opens with a general description of the semantics of the suffix (section 5.3.1). Since my data on verbs of killing do not show all the possibilities mentioned, additional examples will be used. Next, I demonstrate how the different meanings of the suffix are conditioned by the transitive and ergative paradigms (section 5.3.2). Here also, the discussion is not restricted to my database of expressions of killing, but has been expanded to the vast majority of -er nominals in the WSJ corpus, in total more than 45,000 attestations based on 1757 different verbs. While this loosens the ties with the lexical field of my study, the expansion is necessary to rebut the objection that the results obtained from my database would be unreliable because based on only a small number of verbs (some 40 odd verbs if other related verbs like eliminate, slash, exterminate, decimate etc. are included). As the discussion will show, the enlarged data set does indeed confirm the findings for the smaller set of my tagged corpus of expressions of killing. In any event, illustrations will be drawn from the field of killing as much as possible. 5.3.1 The semantic network for -er Rappaport, Malka Hovav Ryder, Mary E."Two recent studies by Ryder (1991a,b) provide excellent insights into the semantics of the -er suffix. By adopting a semantic point of view, Ryder is able to account for the diverse usages of-er in a more systematic manner than does the predominantly syntactic approach of Levin & Rappaport (1988). One of the major shortcomings of the latter is its appeal to an ill-defined distinction between 'agentive' and 'nonagentive' verb arguments. Furthermore, following Wojcik (1976) and Marantz (1984), Levin & Rappaport distinguish between 'intermediary' and 'facilitating or enabling' instruments. They claim that only the former, which can also surface as clausal subjects, can be expressed by an -er nominal. Thus, one can say (can-)opener because a construction like The new gadget opened the can is acceptable. By contrast, one cannot felicitously use *eater in reference to the fork with which one eats, due to the ungrammaticality of *The fork ate the meat. However, as I have argued above, this line of reasoning is too ontological


129

and does not take into account the cognitive structure of grammatical categories. Ryder's account, in contrast, accords well with the properties the Agent exhibits as a prototype-based category. She posits that the prototype of the -er suffix, both diachronically and synchronically, is its reference to human agents, as in Old English writere ("writer"). From this prototype, -er extended to coding the instrument (e.g. clipper, cleaver, screwdriver, computer, etc.). This extension is not unique to the -er suffix, but, as will be recalled, occurs at clause level as well, when instruments surface as clausal subjects. Recall that the extension to instruments is the result of a reinterpretation that pushes the ultimate human agent outside the boundaries of the episode, promoting the instrument to the most agent-like participant. Interestingly, Ryder shows that for the -er suffix, the extension to instruments appears to have been motivated originally in the period from the 1500's through the 1800's, a period of great proliferation in implements and machines whose operation became increasingly independent of human operators. This aligns with the idea already mentioned earlier that the indepen dence of operation of automata is also the locus of an ergativization process characteristic of Modern English. It can be noted in passing that the earliest OED attestations of kill with an inanimate Actor, e.g. the new waggon [is] a mule killer start to occur precisely in that period, which gives further support to Ryder's claim. In the 19th century, there is also a great influx of usage οf killer in reference to chemicals that kill weeds and/or pests, an evolution which finds a parallel in the massive influx of -cide formations referring to such chemicals (e.g. fungicide, herbicide, etc.). With this morpheme the shift is also from 'agent' to 'instrument' (see Cumps 1990 for an interesting discussion). While the use of instrument -er's may have been motivated by the increase of 'selfoperating' devices, the -er formation was reanalyzed to include any type of instrument. As Levin & Rappaport (1988) correctly observe, in present-day English, the largest class of-er nominals refers to instruments. From the instrumental -er another extension, somewhat unusual and apparently not very productive, has developed via a second episode restruc turing, viz. a shift from the instrument used in the episode to the clothing designed to be worn while performing the action. The restructuring has resulted in -er formations like loafers, sneakers, loungers, romper, etc. Significantly in accordance with my earlier description of event Agents, the -er forms have also extended to refer to events that bring about a certain result. Consider for instance, thriller (a film or book that thrills the viewer/reader), bummer ("event that bums a person out") or, within my field of study: (56)

If the show ain't a killer, they're gone. (WSJ)

130


The kill-attestations from the OED on CD-ROM (3753 in total) once more confirm Ryder's view that event -er nominals are a late formation. The only usage of killer interpretable as referring to a causing event is from 1979: that the band was going to deliver another killer set was evident. Ryder points out that in present-day English, this extension goes as far as designating events which are not causative at all. Consider for instance the following forms attested in the WSJ corpus: (57) (58)

Mr. Treiger remembers 90-hour workweeks, including some allnighters (WSJ) It's a no-brainer: You don't have to worry about diversification; you don't have to worry about quality (WSJ)

An all-nighter is an event that lasts all night; a no-brainer is a move or decision that requires no brain to perform. Ryder (1991a) has attested numerous coinages of this type in sports commentaries. More recently, Ryder (1997) has made the claim that the increasing productivity of these formations suggests a process of on-going grammaticalization in which the semantics of -er is bleached, i.e. it is gradually loosing its agentive (causative) meaning. It has become clear that the semantic network Ryder proposes for the -er suffix conforms to the pattern that has been shown for types of Agents that can occur with a wide range of action verbs. The common definition that the suffix designates the agent of the base verb is still valid, provided that one realizes that the notion 'agentive' is a prototypical one covering not only the typical human volitional agent, but also instruments, and events. Interestingly, differences in salience of either 'agent' or 'instrument' give rise to ambiguities with -er formations. For example, a grinder prototypically evokes the conception of the instrument that grinds, but the agentive reading, i.e. interpreting it as profiling the person who grinds (professionally), although less prominent, is not excluded. However, as Levin & Rappaport (1988) point out, the larger construction in which the agentive nominal occurs often helps to disambiguate it. Consider their contrastive pair for grinder: (59) (60)

You find a grinder of fine coffees around the corner. (Levin Sc Rappaport 1988) You find a {coffee-grinder/a grinder for coffee} around the corner. (Levin & Rappaport 1988)

In (59), grinder can only be interpreted to refer to the person who grinds coffee habitually or professionally (ignoring the potential metonymical extension to the shop). In (60) it definitely refers to the machine that grinds coffee. Perhaps a more common example than Levin & Rappaport's grinder is cleaner which can


131

refer either to the person doing the cleaning ('agent') or to the instrument or product used for cleaning. In the WSJ corpus, for instance, cleaner refers in 43% of the cases to a machine used for cleaning (mostly, a vacuum cleaner), in 39% to a chemical substance to clean with, and in 17% to the person who does the cleaning. The context suffices to disambiguate the different usages. 5.3.2 A lexical-paradigmatic view on -er Thus far, the description of the -er suffix has avoided the issue of how the meaning and the distribution of the agentive suffix are conditioned by the transitive and ergative paradigms. The discussion has focused on the more schematic category Agent, which schematizes over the transitive Actor and the ergative Instigator. The restriction to the category Agent is justified in view of the fact that the derivation of agentive nominals is to be situated at this more schematic level. An -er nominal designates the Actor, when the underlying construal is transitive, and the Instigator, when the underlying construal is ergative. For example, a killer is someone (or something) that kills and a choker is someone (or something) that chokes somebody else. Yet the story of the distribution of the-er suffix is more complex than that, as the paradigmatic influence makes itself felt in a variety of ways. 5.3.2.1 Transitive preference of -er. To begin with, in view of the characterization of the meaning of the -er suffix as profiling the Agent, it is to be expected that this type of nominalization will be far more frequent with transitive verbs, which by definition are Agent-centred. Ergative verbs, which gravitate toward the Medium, are expected to be less inclined to shift the profile to the Instigator via agentive nominalization. My (extended) corpus of verbs of killing undeniably confirms this expectation: -er derivation occurs only with transitive verbs, as indicated in Table 5: LEXICAL ITEM

killer murderer shooter butcher executioner exterminator slasher strangler TOTAL

FREQ

PCT

68 25 13 12 3 2 2 1 126

54.0% 19.8% 10.3% 9.5% 2.4% 1.6% 1.6% 0.8% 100.0%

Table 5: Distribution of agentive nominals

132


It should be recalled that the agentive noun butcher is the etymological basis for the verb and is as such a non-typical instance of the -er derivation. Interestingly, however, the OED mentions the double formation butcherer (e.g. Nero, that inhuman Butcher er of Christians, OED 1689), attested only in reference to the brutal killing of people. The formation may well have been the result of a desire to distinguish the wrong-doer from a 'regular' butcher. Table 5 confirms, at least within the field of killing, the Agent-centredness of transitive verbs. Ergative verbs are less concerned with profiling the Instigator in this way, as reflected in the total absence of -er nominalization in the WSJ corpus. In the additional contemporary corpora consulted, there are also no -er formations with ergative verbs of killing. These formations are of course possible, yet less common. The OED mentions citations for most of them, except for suffocate. This may be an accidental gap due to loss of historical documents. To further test the results suggested by my data, I have set up a pilot study of more than 45,000 -er nominals from the WSJ corpus. The selection conditions have been as follows. First, I have retained only -er forms and disregarded semantically related derivations in -or (e.g. professor), -eur (e.g. entrepreneur), -ier (e.g. bombardier), etc. since they are mostly loanwords from French or Latin. Secondly, I have retained only -er nominals which are derived from verbs. Excluded are thus nominals like one-liner, islander, jeweler, etc. In some cases, I did retain formations which are only indirectly derived from verbs. Mostly these concerned agentive nominals derived from compounds or derivations like onlooker, frontfighter, roadrunner for which there is no base verb *to onlook, *to roadfight, *to roadrun. Given the salience of the uncompounded base verb (look, fight, run), these formations have been included. Note that sometimes such compounded base forms may conventionalize as a verb, e.g. brainstormer (to brainstorm), spellchecker (to spellcheck). (The correct chronological development of such formations is not directly relevant to the present study.) Very often, these complex agentive nominals incorporate the Affected as first element of the compound, as already illustrated by coffee-grinder above. Here are some more suggestive WSJ-examples: make: lawmaker, homemaker, automaker, shoemaker, tire-maker, filmmaker, rainmaker, decisionmaker take: risk-taker, hostage-taker, test-taker, picture-taker, poll-taker, care-taker build: ship-builder, homebuilder, bomb-builder, bodybuilder kill: painkiller, weed-killer, copkiller break: circuit-breaker, ice-breaker, heartbreaker, lawbreaker, groundbreaker, quota-breaker


133

Another restriction to the selection of -er formations has been the exclusion of proper nouns, like Baker, Fisher, etc. Although etymologically these names derive from agentive nominals, they no longer activate the agentive component. Finally, a major restriction concerns the source used: for practical reasons, the forms have been extracted from the frequency list of all words in the WSJ corpus, but I have not systematically traced the context and/or meanings of each attested nominal. The overwhelming amount of data (45,600 sentences to be examined and coded) clearly exceeds the purposes of the present investigation. Nevertheless, I believe that my pilot study is relevant, all the more so because of the high number of compounds like the ones listed above which provide clues for correct analysis. For some typically ergative verbs, like break or open, I have examined the actual attestations. Despite a considerable degree of accuracy, the results presented in the following part should be approached with a certain amount of caution. For each extracted agentive nominal, I have classified the base verb as essentially transitive or ergative. Verbs that show mixed characteristics, like sell, have been marked accordingly. The result of the tabulation is presented in Table 6, which shows that more than three fourths of -er normalizations are based on transitive verbs, only 2.5% are based on ergative verbs. VERB TYPE

FREQ

PCT

transitive intransitive ergative transitive & ergative

35395 8065 1182 974 45616

77.6% 17.7% 2.6% 2.1% 100.0%

TOTAL

Table 6: Overview of verb types with -er nominalization As such, the table confirms the results for my own analysis of expressions of killing, for which -er derivations occurred only with transitives. Yet, there is a serious methodological complication to Table 6, since the results may say more about the ratio between transitive and ergative verbs in the WSJ in general rather than about the distribution of the -er suffix in particular. That is to say, if the distribution of transitive, intransitive and ergative verbs is approximately 78%, 18% and 3% respectively (ignoring the blends), then the results of Table 6 are merely a confirmation of this general distribution and should not be interpreted as signaling the markedly transitive character of the -er suffix. This poses a formidable problem for my investigation, since it requires that I determine the general distribution in the WSJ corpus of intransitive, transitive and ergative verbs before I can interpret the results for the

134

LEXICAL PERSPECTIVES ON TRANSITIVITY AND ERGATI VITY

-er suffix. This task is not feasible at present, which puts an unfortunate practical limit on the depth of my inquiry. - However, a possible solution to the methodological problem is to examine the distribution in a subset of the WSJ corpus. My coded corpus of expressions of killing and dying can serve as a basis for estimating the transitive/ergative ratio in the entiîe WSJ corpus. The results of the computation is presented in Table 7: VERB TYPE

transitive intransitive ergative transitive & ergative TOTAL

FREQ

PCT

1121 556 130 43 1850

60.6% 30.1% 7.0% 2.3% 100.0%

Table 7: Overview of verb types with verbs of killing The ratio of intransitives is higher than perhaps expected, because items like die or perish have been included in my corpus for contrastive purposes. The ergative verbs are still a definite minority (7%), but score better than in the erbased tabulation. Given the marked transitive character of the selected field, I expect that the general ratio of transitive and ergative verbs will approximate 9 to 1. However, despite being suggestive, a comparison with a diversion of 4.5% (the 2.5% of ergative verbs with -er in the entire WSJ versus the 7% ergatives in my corpus) does not justify any definitive conclusions. To investigate the distribution of the -er derivation more systematically, an all-encompassing analysis, like the one I have conducted for the field of killing, should be applied to a much more extensive and variegated corpus. The WSJ corpus may be somewhat one-sided showing some frequency peaks for particular coinages (e.g. maker and its compounds add up to 3313, almost 10% of all transitives). To sum up, while indicative a mere frequency analysis of verb types cannot unequivocally determine the transitive nature of the -er suffix. Another possible perspective on the agentive suffix, however, is to consider the derivation as an inherently transitivizing one. I have already emphasized that the -er suffix indeed derives an agentive nominal, as it profiles the Agent of the base process. What is taken as Agent is a matter of conceptualization, as manifested in the extensions from Agent -er to Instrument -er and Event -er, which agree with those possible with the base verbs. The suffix thus inherently shifts the focus to the Agent or to the action component, even with derivations from ergative processes. However, things are more complex, as there are two paradigmatic complexities are to be considered, viz. Medium-profiling 'agentive' nominals and middle-based derivations.


135

5.3.2.2 Paradigmatic complexities of -er. Keyser & Roeper (1984) argue at length that when affixing to ergative verbs, -er can only refer to the Instigator, never to the Medium: "Sharpener may not refer to the debate in The debate sharpened, nor may dropper refer to the icicles in The icicles dropped" (1984:395). This forms a major part of their argument, as they claim that the -er suffix is in complementary distribution with the prefix re-. The latter occurs with transitives (e.g. repaint) and ergatives (e.g. reopen, reharden), but not with 'true' intransitives (e.g. *resqueak). Keyser & Roeper explain the occurrence of the "transitivizing" prefix re- with ergatives by interpreting these verbs as "intransitives whose subject has been moved from object position" leaving an object trace. This allows them to say that re- is in a sense always transitive. Since in their view the suffix -er is not linked to a trace but to a full NP subject, it cannot occur with ergatives in reference to the Medium, as the following selection from their examples illustrates: (61)

reorganize reaccelerate readjourn reharden

*organizer *accelerator *adjourner *hardener

(= the team) (= the car) (= the meeting) (= the metal)

Although true of the vast majority of ergative verbs, their account is not entirely correct, since there are clearly -er derivations that profile the Medium. Further more, there are also transitive based -er formations that profile the Goal. These two complexities are taken up in turn. Medium-profiling derivations. For -er derivations that profile the Medium consider, for instance, cracker. In one reading a cracker is not someone or something with which you crack objects (e.g. a nutcracker), but a crisp thin wafer that cracks when you eat it. A circuit-breaker is undeniably a device that 'breaks' the electricity circuit (or some other 'circuit', e.g. that of stock-trading) and thus profiles the Instigator. Yet breaker can refer to (the part of) a wave that breaks itself into foam on the shoreline, a conceptualization in which the wave is clearly the Medium of the underlying break-process. Within my field of study, the OED lists the following usages οf starver and drowner. (62)a. b. (63)a. b.

... corn-dealers are starvers of the poor... (OED, 1859) So long I have wander'd a starver ...(OED, 1839) Ydlenesse, enemy of vertue, ye drowner of youthe (OED, 1545) As waves ... Hiss round a drowner's head. (OED, 1820)

In the (a)-sentences, the -er nominal refers to the Instigator, in the (b)sentences, to the Medium. Unfortunately, I have no attestations from any of the

136


contemporary corpora, yet it is intuitively plausible that the -er suffix may come to profile the Medium with these verbs. If people were asked what drowners or starvers are, some might certainly be inclined to identify them as Mediums, not Instigators. The Medium's considerable independence in the designated processes is responsible for the somewhat unexpected Medium-profiling -er derivation. In the case of starvation and drowning in particular it is really the Medium that is 'doing' the process, while the Instigator remains quite external to it (cf. Chapters 4 and 6). In such cases, the -er suffix can profile the Medium, the entity which is the most salient agentive participant. It should be observed, however, that the profiling of the Medium has not been attested with the other ergative SUFFOCATE verbs, like choke or stifle. This is so because they incorporate a more salient reference to an Instigator. While it is true that agentive nominals derived from ergative verbs profile the Instigator in the vast majority of the cases, Medium profiling -er forms must not be dismissed as "idiosyncratic, noncompositional meanings" (Keyser & Roeper 1984:397). Although somewhat unpredictable, these formations are motivated by the stronger independence of the Medium. The higher the independence, the more likely it becomes that an -er formation may profile this participant rather than the Instigator. In this context, it should be pointed out that Medium profiling agentive nominals often occur in the transition area between transitives and ergative noneffectives. As has been illustrated in Chapter 4, section 4.3.1, this area has been the locus of important linguistic change in the past centuries, as many initially purely intransitive processes have become representable as 'instigatable', tilting their clause constellations towards the ergative paradigm. It will be recalled that also starve has retained a strong focus on the participant that starves, the former intransitive Actor, whose salience triggers the possibility of a Medium profiling -er formation. In addition to starve, also verbs like stick or float can be mentioned here. These are basically intransitive verbs that can receive a forcedynamic coding, as in She stuck a note on my door or They are floating the logs down the river. Given the intransitive basis,s t i c k e r sandfloatershave conventionalized as profiling the originally intransitive Actors and generally do not refer to the Instigators of the sticking and floating processes. So, instead of saying, as Keyser & Roeper (1984:395) do, that ergative noneffectives "are open to reanalysis as [intransitives] that allow -er", I am inclined to argue on the basis of my preliminary investigation of -er nominals, that in general the reverse is the case, viz. that intransitives that allow -er have become ergativized by extending leftward to include the instigation. But what about the ergative verbs? As a rule, they align with the transitives as -er derivations of ergative verbs profile the Agent. The occasional Medium-profiling derivations, like breaker and starver, are motivated by the salience of the Medium. It is the


137

semantics of the usage that is responsible for the salience of the self-instigating participant. Often, the Instigator is very much backgrounded in these usages; asking who or what is the Instigator of waves breaking on the shoreline is an odd question. In these uses, the Medium is almost promoted to an intransitive Actor and Keyser & Roeper's view of a reinterpretation of the ergative as an intransitive may apply here. Thus, the -er derivation seems to imply a strong pull towards the actional mode. Instead of assuming as we did earlier that -er nominalization is predominantly restricted to transitive verbs, would it not be more correct to say that the derivation is a transitivizing one by its strong profile on the Agent or by its promotion of the Medium to (intransitive) Actor? The inaccuracy of this deduction is illustrated in the next part. Goal-profiling derviations. Saying that the -er nominal is inherently a transitivizing one is incorrect, as the derivation does not always code an active participant. Important evidence comes from yet another paradigmatically determined construal, viz. agentive nominals based on middle usages of transitive verbs. Typical cases are the following examples cited both by Ryder (1991b) and Levin & Rappaport (1988): (64)

dipper: something that is dipped before being eaten sipper: a drink that is sipped scratcher: a lottery ticket to be scratched to reveal the potentially winning pattern

I believe that Ryder (1991b: 309) overextends the agentivity implied by the -er suffix in her characterization of these formations as profiling an "agentive patient", since these formations are the nominal equivalent of a middle construction, which we argued to profile a Goal (Chapter 4, section 4.2). Although these formations profile the Goal of the transitive base process, the -er formation does not promote this participant to Agent. I argue that middle-based -er formations like those in (64) retain the same voice constellation: the essence of the entities referred to is to be dipped, sipped or scratched by some implied agent. So, as with the middle construction the focus is on the properties of the Affected. Consequently, and in line with our account of the middle construction, such a middle-based derivation will generally not occur with verbs whose semantic structure is incompatible with this focus on the properties of the Affected. Agentive derivations of verbs of killing, which were shown to resist the middle alternation, such as killer or murderer indeed saliently refer to the Agent. Nevertheless, like the middle alternation, a middle-based -er derivation can occur in the proper context. Consider two examples with killer occurring in the

138


same context as the earlier mentioned unexpected middle constructions, viz. in reference to cattle ready or predestined for slaughter: (65) (66)

...the bullock among the herd of killers ... (OED on CD-ROM) As killing cattle or killers (cattle ready for killing) they are inferior to corn-fed stock. (OED on CD-ROM)

Although unprototypical, these middle-based uses of killer, once more, signal that the type of construction a verb can engage in is determined by specific usages and cannot be stated in absolute terms. Further, in addition to the other examples of Goal-profiling nominals, these usages reveal that middle-based -er formations differ slightly in meaning from the regular middle. As explained in Chapter 4, the middle construction prototypically focuses on the processfacilitating (or process-hindering) properties of the Goal. Similar to what was observed for atypical middle constructions with verbs of killing, the Goalprofiling -er formations shift the focus to the Goal's essential destination. A dipper is not a snack that has properties that facilitate dipping it (which would distinguish it from other snacks), but one that is meant to be dipped. Similarly, a scratcher is a ticket meant and designed to be scratched. Similarly, the cattle referred to as killers are fattened for slaughter and thus predestined for it; the formation is not to be interpreted as indicating that these animals possess qualities that facilitate the actual killing. In short, these middle-based derivations specify that defining feature of the Goal is precisely to be subjected to the process denoted by the base verb. In the discussion of the middle construction in Chapter 4, it was also observed that middle constructions can give rise to blends, i.e. verbs that show both transitive and ergative characteristics. Most interestingly, these blends find their counterpart in the -er formations as well, as indicated by the following contrastive pair: (67)a. bookseller: person or store that sells books b. bestseller: book (or possibly another product) that sells well (not: book that is sold well) A best-seller's properties are such that selling it is no problem; in fact, it sells itself. The latter construction clearly entails an ergative reinterpretation of an originally transitive process. The WSJ corpus, which often reports on buying and selling, is well-suited to do a pilot study probing into the distribution of the two types of -er's for seller. On a total of 351 attestations of seller, the vast majority (86%) profile the Agent. Of the Affected-profiling formations (14%), the majority (73%) concern attestations of bestseller. Strikingly, all the nominals profiling the Affected retain a strong facility-oriented perspective that


139

typifies the middle construction: they all contain a modal comment on the selling process. In addition to bestseller, we find big(gest) seller, hot seller, top seller, slow seller, third-largest seller, and huge seller. This reveals the item's transitive homeground and confirms the prominent feature of modal comment on the feasibility of the selling. In the discussion of the middle construction it was pointed out that this also holds on clause level: in spite of an adverbless clause construction that is possible with sell (e.g. this magazine sells), the clause retains its focus on the feasibility-enhancing properties of the object sold. In other words, even adverbless constructions with sell still very often imply a comment on the product's properties. This reveals that while blends like sell or wash or drive are both transitive and ergative, the transitive paradigm still predominates. The predominance of the transitive paradigm is also reflected in the fact that -er derivations do not always inherit the mixed pattern: with considerable certainty, one may say that washer and driver will always be used unambiguously in reference to the Agent of the washing and driving processes. This is also the case for agentive nominals derived from transitive verbs of killing. Despite the two OED examples, formations like killer or murderer saliently refer to the Actor of the base process. 5.3.3 Summary In conclusion to this discussion of -er formation, it can be noted that in this area of verb-based morphology, the transitive and the ergative paradigms are also operative. It has been revealed that the suffix prototypically profiles the Agent, which correctly predicts that it will more readily apply to transitive than to ergative verbs. In the case of some ergative verbs, an agentive derivation may be marginal or excluded altogether (strikingly, for example, suffocator is absent from all the corpora consulted). On the other hand, the strong independence of certain ergative MEDIUM-PROCESS constellations—inherited from the original intransitive construction that has been subject to ergativization—motivates an extension of the -er suffix to profiling the Medium instead of the Instigator. For other 'true' ergative based -er derivations that profile the Medium, the semantics of the particular usage motivates the unusual profile. Parallel to the recent development of middle constructions within the grammar of transitive processes, a most recent evolution in the -er network (and one that is still fairly uncommon, cf. Ryder 1991b) is the use of the -er nominal in reference to the Goal of the transitive base process. As with middle constructions, the profiled entity does not become an Agent but remains an inert Goal which possesses certain characteristics that almost predetermine it for being 'V-ed'. Clearly, my investigation into both the distribution and the meaning of agentive nominals has only scratched the surface of what appears a complex phenomenon in the grammar of English. Despite some undeniable tendencies

140


that have been revealed by the above discussion, a more all-encompassing research is obviously warranted. My pilot studies of the -er nominals in the WSJ corpus and that for seller indicate the avenues along which such a research may be conducted. The first pilot study has shown that in order to come to valid conclusions concerning the 'ecological niche' of the -er suffix, i.e. the transitive paradigm, it is essential that one considers the overall distribution of the transitive and ergative paradigms (which I assume to take the form of a 9/1 ratio). Only then can one arrive at a proper evaluation of the distribution of-er. The WSJ-based investigation into the type of participant profiled by seller reveals moreover that a more detailed analysis is indispensable if one wants to give a coherent account of the semantic extensions of the agentive nominal. In the concluding chapter to this book, some avenues of further research will be discussed more elaborately. 5.4 Constructional blueprints f or agentivity: the objectless transitive 5.4.1 A lexical view on the objectless transitive In the preceding sections, the description of the Agent-centredness as an essential feature of the transitive paradigm has proceeded from a lexical and morphological perspective. The present section (which like the previous is less saliently tied to verbs of killing) looks at the Agent-centredness from the constructional end. More specifically, I will consider the objectless transitive, a construction which maximizes the Action component. In this type of construction, a transitive verb occurs without an overt Goal, as in Soldiers kill, John is eating, etc. Characterizing these structures is a challenging problem, since they entail a shift from a 'transitive' process to an 'intransitive' one.48 In the grammar of English, a similar type of alternation (viz. the alternation between a formally 'transitive' and formally 'intransitive' verb construction) involves the ergative effective/non-effective alternation (as in John starved Mary versus Mary starved). Can the objectless transitive then be defined as the transitive equivalent of ergative non-effectives? In a sense, such a definition is correct, since both construction types share the feature of isolating the participant that is structurally and conceptually central to the transitive and ergative paradigms, the Actor and the Medium respectively. Nevertheless, there are important differences, since the objectless transitive still implies the other participant, the Goal, whereas the ergative non-effective remains neutral on the question of the instigation: the process can be either self-instigated or instigated by an external 48

See also Allerton 1975; Fillmore 1986; Fellbaum & Kegl 1989; Mittwoch 1971, 1982. Jespersen 1927 presents an impressive overview of 'transitive' constructions used 'intransitively'.


141

force. Given the inherently Goal-directed nature of the objectless transitive, it does not classify as an intransitive either; it is essentially an effective constellation (cf. Davidse 1991:23). A similar claim is forwarded by Rice who observes "the object does not really go away when it is omitted", meaning that "they are still present at some level of organization, perhaps not at a lexical or syntactic level, but certainly at a conceptual one" (1988:203)49 If the objectless transitive is inherently effective, then the question is what distinguishes it from transitive constructions that are overtly effective, i.e. those that do express a Goal. What is it exactly that motivates an objectless construction? In her quest for an answer, Rice starts from the general idea that "objects that can be omitted tend to be those whose lexical content is most probable given the meaning of the verb" (1988:203). She looks at the problem from a lexical perspective and tries to prove that the degree of specificity of both verb and object determines the possibility of object omission. Very general verbs that sustain a wide variety of complements tend to require objects. Also verbs that are quite specific with regard to their object or reveal something about the manner in which the action is carried out (e.g. John puckered his lips or Mary manicured her nails) almost always require overt objects. Thus, Rice says, object omission occurs mostly when the verb is moderately neutral and the range of objects is restricted to one or two possible semantic domains. An objectless construction like Thou shalt not kill is a case in point: the verb is relatively neutral concerning the specifics of the event and the Goal is relatively schematic and can be interpreted minimally as human beings or maximally as living beings. Rice concludes that "omissibility [...] can be motivated for classes or strings of verbs rather than for idiosyncratic lexical items" (1988:207). 5.4.2 A lexical-paradigmatic view on the objectless transitive Rice's account goes a long way towards explaining what motivates object omission with transitive verbs, yet still falls short of providing a coherent account, because she looks at the phenomenon exclusively from the lexical end. In line with my findings on the interaction between lexical and constructional meaning, I argue that Rice's lexically inspired point of view should be augmented with a constructional perspective. This would allow her to account for some unexpected exceptions to her generalizations. First, Rice does not distinguish between ergative and transitive verbs. Consequently she overlooks an important generalization, viz. the impossibility 49

Fillmore (1986:95) notes that such verbs are "verbs which can occur either with or without a direct object, but which are understood as having, when used intransitively, an understood object". Bolinger (1975:209) makes a distinction between "goal-transitives" (e.g. John lost the money) and "goal-intransitives" (e.g. John lost).

142


of an objectless construction with ergative verbs, e.g. *someone opened/closed is not a valid alternation to someone opened/closed the door. The unacceptability is not to be accounted for on grounds of the broad range of possible objects which disfavours omission; it is determined paradigmatically: oneparticipant constructions with ergative verbs can only be interpreted as Medium-Process constellations. (Hence, the term 'objectless' always refers to a transitive construction, never to an ergative one.) Secondly, the feasibility of an objectless construal is context-sensitive. For instance, an objectless construction God saves is quite natural, but *This lifesaver saved is clearly infelicitous. Alternatively, as Rice herself notes, transitive verbs that usually do not allow an objectless construction may occur in contexts in which omission of the Goal is possible, despite a wide range of object possibilities: (68) (69) (70) (71)

Hemingway ate, drank, and smoked too much. (Rice 1988) Scott hammers and saws like a pro. (Rice 1988) Bill always interrupts. (Rice 1988) The freedom-fighting contra rebels kidnap, rape, torture, and murder. (Rice 1988)

Note that make, which Rice explicitly identifies as a verb not sanctioning an objectless construction, does allow it. Consider, for instance, many cases where objectless make occurs with particles or prepositional complements, as in She made after him, Everyone made for the bar, The car made off at top speed, or How are you making out with Lucy? Another fine example is the phrase Man makes, God creates (Kristin Davidse, pc).50 To account for (68)-(71), Rice observes that for the events denoted by these examples, "the particular object is fairly unimportant as the pragmatic focus is on the activity itself" (1988:206). Within a particular semantic frame (e.g. that of war and/or (organized) violence underlying (71)), it is possible to identify the omitted object. I argue that this 'pragmatic focus on the activity' is pivotal to the meaning of the objectless transitive. Hence, it comes as no surprise that, as a rule, they function as general statements, not tied to any specific event. Very often, the general statement is accompanied by a modal comment, like too much, like a pro, and always in Rice's examples cited above. Also negation is common, e.g. Islam does not teach kidnapping and killing (WSJ).51 For this reason, it can be 50

Of note are also some (slightly old-fashioned) phrases such as to make merry/bold/free (Jespersen 1927:331) or to make ready. 51 The gerundive construction is also taken as a valid instantiation of an absolute construction since the nominalization applies only after the direct object has been integrated (in this case, omitted). This is what distinguishes gerunds from the -ing nominalization that applies at a


143

claimed that, in a sense, the objectless transitive is the 'Actor-focusing' counterpart of the 'Goal-focusing' middle construction, as both constructions are not tied to individuated events but imply a general comment on the process. In essence, then, the objectless transitive focuses on the general characteristics of the act and cannot, in principle, refer exclusively to a specific event. Even when tied to a specific act, the objectless transitive positions the act in a more general perspective. Consider two of Rice's examples: (72) (73)

Martha cooked and cleaned while Mary entertained. Billy Jo washed and Bobby Jo dried.

These sentences clearly refer to specific individual events. Nevertheless, the objectless constructions still have a generalizing effect, taking the participants' actions to the more general plane of role (stereo)types within a given frame (for (72) that of receiving company and for (73) that of washing up). Moreover, each time, the 'role stereotype' of the one participant is contrasted to that of the other. When the latter is omitted, the acceptability of the objectless construction decreases considerably: (72') ?Martha cooked and cleaned. (73') ?Billy Jo washed. The reason is that the generality of the statement is reduced. Note, however that these usages can be reinterpreted within some other general frame, as in Martha cooked and cleaned there for years, On Fridays, Billy Jo washes, or When I am on holiday, I do not cook. The earlier quoted example Man makes, God creates, similarly opposes two roles, which sanctions the objectless construction. Recall also the LDC example old enough to bleed, old enough to butcher, which is the only absolute construction for butcher in any of the text corpora. The general character of the objectless construction also explains why the verb in these constructions very often occurs in the simple present tense and less often, for instance, in the progressive, although the latter is not impossible (e.g. Ford is buying at $110 a share). Significantly, none of Rice's acceptable examples of the objectless construction occur in the progressive. My corpus of killing-expressions does not allow us to investigate the properties of the objectless transitive construction more systematically, since the construction does not occur frequently. The objectless construction is in principle possible with all MURDER verbs (see, for instance, example (71) lower level, i.e. when -ing attaches to the base verb. In that case, the Agent and Affected can only be expressed periphrastically (e.g. the killing of Higgins by Shiite terrorists). I have not interpreted omission of the Goal in these nominalizations as an absolute.

144


above), although intuitively it is less likely to occur with verbs like massacre and assassinate: (74) (75)

??The government troops massacred. ??The cartel assassinates.

Possibly, the stronger Goal-orientedness implied by these predications is a constraining factor. In line with my characterization of the meaning of the objectless transitive, the attested objectless constructions all generalize over different events or are statements concerning the act of killing itself, independent of the particular (type of) Goal at which it is targeted. Here are some typical examples: (76) (77)

... the current James Bond film, "Licence to Kill," ... (WSJ) Director Roger Spottiswoode, who directed "Shoot to Kill"... (WSJ)

(78)

... the anguished serial killer turns himself in to police and says, "Stop me before I kill again." (WSJ)

In the context of the James Bond film, the omitted Goal is self-evidently interpreted as being a human being. It would be awkward if, say, James Bond were to justify his shooting of a dog by claiming that, as secret agent 007, he has the license to kill. If one has the right to kill humans then one consequently has all the other rights and there is no need for (explicit) justification. A hunter of deer, in contrast, would not claim that he has 'a license to kill' (but rather say he has 'a license to kill, hunt, or shoot deer'). The variability of interpretation shows that, in fact, the true nature of the Goal is irrelevant, since it is the act itself to which the license applies. Similarly, the last example about the serial killer focuses on the (repetitious) act of killing, but this time the focus is not on the licensing but on the illegitimacy of the act. The earlier cited commandment Thou shalt not kill equally makes a statement about the inappropriateness of the act. An interesting example is the following: (79)

Most of the millions that perished at Auschwitz never even slept there for a night. They went straight from trains to 'showers' under Zyklon Β, which killed in minutes... (WSJ)

The focus on the 'mad efficiency' of the poisonous gas has most likely triggered the coding of the gas as Agent (cf. section 5.1); most probably, the consequent focus on the Action-component has also invited an objectless construction. The objectless construction further emphasizes the insignificance of the Goal in the massive killing reported on in this sentence. The sentence shows that, in general, the objectless transitive is not restricted to animate or volitional Agents. A case


145

in point is the slogan of the U.S. pro-gun lobby It's not guns that kill (people); often the objectless occurs next to a regular effective. In this case, the focus is clearly on who or what is responsible for the killing. The above discussion has revealed that explaining the objectless construction on the basis of primarily lexical considerations tells only part of the story. It is essential to a proper characterization of the construction that one also considers the semantics of the construction itself. Such an integrated perspective indicates that the general character of the verb-object pair is not really, as Rice has it, a strict precondition for the objectless transitive but rather a consequence of the construction's implication of a general statement which abstracts away from individual instantiations. The construction maximizes the focus on the Action-component, which also explains, as Davidse (1991:33) points out, why an objectless transitive construction is impossible with clauses expressing an accidental transitive action: (80) (81)

* John accidentally touched. (John accidentally touched her eye.) *The doctor accidentally hit. (The doctor accidentally hit a nerve.)

The unacceptability of the objectless in these clauses is a natural consequence of the fact that in transitive clauses with non-volitional animate Actors, the semantic focus shifts from the action-component to the goal-achievement component, as was also pointed out by Nishimura (1993). The focus on the goalachievement is incompatible with the maximization of the action-component implied by the objectless transitive. Although transitive clauses with an inanimate Actor also shift the focus to the goal-achievement, they can nevertheless occur in an objectless construction as example (79) shows. As already observed by Davidse, this indicates that "goal-achievement by an inanimate Actor is a closer instantiation of the transitive prototype than accidental goal-achievement by a conscious Actor" (1991:36). Clearly, next to the more general 'grammatical' meaning of the objectless transitive, the properties of individual verbs do have a certain influence on the feasibility of this construction. Rice lexically oriented observations may signal that the objectless construction prototypically occurs with verbs that are "neither extremely schematic nor extremely specific" (1988:207). The high frequency of objectless constructions with kill may be taken as indicative, but more extensive corpus-based research is warranted. Finally, it can be noted that like with all constructional changes, it may happen that for a given verb the objectlesstransitive construction becomes a well-entrenched pattern, even at the expense of the transitive effective. While

146


for the examples discussed here, such evolution has not taken place, Jespersen mentions a few such shifts to full-fledged intransitive verbs, especially with originally reflexive verbs, e.g. to pine (away), from to pine oneself or to recover, from either to recover oneself or to recover health. In many cases, the objectless transitive entrenches for a given usage, e.g. he drinks· (in reference to his being an alcoholic). Further diachronic research will surely prove rewarding to trace such shifts and the conditions under which they occurred. 5.5 Conclusions In conclusion to this chapter, it can be pointed out that the corpus-based descriptions of the MURDER verbs converge to indicate that the grammar of English acknowledges the Agent-centredness of the transitive paradigm in a variety of ways. The lexical semantic description of the MURDER verbs supports the view that transitive verbs centre around a volitional Actor, yet at the same time shows how the more general verb kill is less strongly committed to such an Actor. The high frequency of event Actors, and especially the frequent Agentless passive with the cause surfacing as a setting, has revealed that such so-called 'atypical' Agents become wholly motivated against the background of contextual and experiential features. The description has further clarified the relationship between Agents and Causes, which essentially boils down to a different degree of participation. The more detailed discussion of the derivation of agentive nominals has not only corroborated the view defended earlier that the Agent-category is a prototype-based category, but has also shown that the -er suffix is strongly tied to the transitive paradigm by its inherent tendency to focus on the Actor of the base process. At the same time, I have demonstrated how further paradigmatic complexities are involved in the -er derivation, since the nominal may come to profile the Medium of an ergative process and the Goal of a transitive middle construction. The analysis of the latter confirms the earlier analysis of the middle construction and like the -able derivation is a morphological confirmation of the prototypical non-middable character of verbs of killing. Finally, the objectless transitive has been shown to reflect the Actor-centred character of the transitive paradigm in its maximization of the semantic focus on the Action-component. The description has made it sufficiently clear that the objectless transitive is semantically different from the (more typical) transitive effective constellation (AC-PROC-GO). Nonetheless, it cannot be conflated with 'true' intransitives, since despite the omission of the Goal, they still code a process that is inherently directed towards a Goal. The schematicity of the Goal should not exclusively be accounted for by the lexical properties but can be


147

explained against the background of the non-individuating nature of the construction. In short, the transitive paradigm can safely be defined as fundamentally and prototypically Agent-centred, a feature which maximally distinguishes it from the ergative paradigm that is primarily concerned with the Affected (the Medium). On the basis of the predominantly ergative SUFFOCATE verbs, the following chapter will probe into the Medium-centredness of this paradigm. However, the chapter will already provide descriptions of cases where paradigmatic opposition is less clear-cut. Taking this issue to its logical conclusion, the last chapter of this book examines more elaborately the dynamic interaction between the two paradigms in its diachronic and synchronic analysis of abort.

CHAPTER 6 THE MEDIUM-CENTREDNESS OF THE ERGATIVE PARADIGM

6.1 The experiential basis of ergativity In the description of the field of killing in Chapter 3 the subgroup of SUFFOCATE verbs was characterized as predominantly an ergative one, as instantiated by the verbs suffocate, asphyxiate, choke, drown, and, less so, stifle. The present chapter probes into this ergative predilection, showing how it is experientially well-motivated as is also the distinction between the transitive and ergative verbs within the group. Diachronic and synchronic explorations into the verbs' lexical and constructional behaviour will reveal an intricate yet highly dynamic network underlying the group of SUFFOCATE verbs. An essential characteristic of the SUFFOCATE verbs is their salient reference to the respiratory system. Such a focus on a particular body part is surely not exclusive to the SUFFOCATE verbs, but occurs with other verbs of killing as well. For example, in the conception designated by the STARVE verbs (starve and famish, and formerly also hunger and thirst), the victim's metabolistic failure (due to lack of nutrition) is quite prominent. For the DECAPITATE verbs, the neck area will be in focus as well as the type of bodily harm, the severage of head and trunk. In fact, within the field of killing it is the manifest non-specificity of the MURDER verbs which is in fact striking. Despite the salience of the respiratory system in the semantic structure of the SUFFOCATE verbs, it is still the victim itself and not the affected part of the respiratory system which is generally profiled as the affected participant. This can be explained drawing on Langacker's notion of active zones. Consider a sentence like John strangled Mary. It profiles a relationship between the two event participants, yet it seems that only selected aspects of the trajector (John) and landmark (Mary) are directly involved in it. More specifically, John's hands (assuming that he used his bare hands as instruments) will be pivotal to the act of strangling; at the other end, the throat is the portion of Mary that is directly involved in the relationship. Langacker (1984, 1993, 1995) has called these focal areas active zones which are secondary profiles within the profiled entity. In fact, they are a special case of metonymy, since one entity (the whole) is invoked as a reference point to establish mental contact with another, the active zone (the part). However, the active zones should not

150


be thought of as discrete regions, but rather as areas which are more prominent than others in their participation in the interaction. Essentially, then, "when an entity is conceived as participating in a given relationship, there is usually some discrepancy between its profile and its active zone for that relation" (Langacker 1995:20, his emph.). The following examples show that such active zone/profile discrepancies are a very natural phenomenon: (1) (2)

Your dog bit my cat. (Langacker 1984) Your dog chased my cat. (Langacker 1984)

Although in example (1) the dog and cat are in profile, only a part of the dog (its teeth, jaws, etc.) and a part of the cat (the part where it is bitten) are directly involved. Through various constructions the active zone can be made explicit, e.g. Your dog bit my cat on the tail with its sharp teeth, in which the first prepositional phrase specifies the active zone of the landmark and the second, that of the trajector. In example (2), in contrast, the active zone and profile coincide and the cat and dog are conceived as undifferentiated wholes. The presence or absence of a discrepancy between the active zone and the profile can be visualized as follows (Langacker 1995:21): (a)

(b)

Fig. 17: AZ/Profile discrepancy in relation predications Figure 17(a) represents example (1): only part of the trajector and landmark are the focal areas directly engaged in the process. The representation in Figure 17(b) corresponds to example (2), for which the whole entities partake in the relation. Langacker claims that a discrepancy between active zone and profile is typical, and situations like in (2) much less common. The commonality of this discrepancy may be the logical exponent of the general cognitive principle that the whole is more salient than its parts (cf. Lakoff 1987). As a result, people will be more inclined to profile the whole rather than the parts. The SUFFOCATE verbs also typically profile the victim as a whole, although profiling the active zone as the Affected is in principle possible: (3) (4)

A hundred times their throats choked as the car careened on a bank. (OED on CD-ROM) ... the dank air and dust choked his lungs (HHGG)

THE MEDIUM-CENTREDNESS OF THE ERGATIVE PARADIGM

151

This type of usage is, however, quite unusual, probably because in the case of a suffocation event the affected body part is less likely to be conceptualized as independent of the body and, more importantly, because the process does not just affect this part but the whole person whose life is at stake. This is a marked difference with some other verbs of bodily harm, which obligatorily or prototypically profile the affected body part, e.g. cut (fingers, limbs), sprain (ankle, wrist, knee), prick (finger). My data reveal that if an explicit reference of the active zone occurs within the group of SUFFOCATE verbs, it is mostly with choke. In addition to examples (3) and (4), consider the explicitness of the two active zones in the following example: (5)

The pirate's hissing [...] was half choked in his windpipe by my clutching fingers. (NOV)

As said, at the most general level, the active zone that differentiates the SUFFOCATE verbs from other verbs of killing is the entire respiratory system. This complex system itself, however, divides into various subparts, viz. (from outside to inside) the mouth and nose, the throat's aperture, the windpipe, and the lungs. The SUFFOCATE verbs are primarily differentiated by the type of cause and a different active zone. The two are related, since a different cause is likely to affect a different part of the respiratory system. Interestingly, the general division into ergative and transitive SUFFOCATE verbs is also tied to a general distinction in active zone. The ergative verbs tend to have as their active zone the most internal parts of the respiratory system (the lungs, the throat's aperture), whereas the transitive verbs have the more external parts (the outside throat) as their active zones. This distinction can be represented as follows52:

internal zone

external zone transitive strangle throttle

transitive-(ergative) smother

stifle

ergative drown

suffocate

choke

asphyxiate

Fig. 18: Internal/external and ergative/transitive correlation Although this distinction does not represent a rigid dichotomy, it gives a first assessment of the experiential basis of the ergative preference of the group as a whole as well as the transitive character of some of its members. That is, the 52

The metaphorical and ergativized extension throttle an engine is not included as it is not relevant here.

152


ergative conception seems to be strongly conditioned by the ontological characteristics of a suffocation process, for a number of reasons. First, the process involves a considerable degree of bodily activity on behalf of the suffocating entity; it is in fact the victim that is 'doing' the suffocation. Clearly, the participation is not deliberate yet it is unavoidable since the physiological processes affect the victim's respiratory system. Secondly, the causes of suffocation events, although multifarious, are typically imperceptible, as in the case of gasses or (lung) diseases. This imperceptibility encourages a conception of the suffocation as independent of the cause or instigation, which is an essential feature of the ergative construal. Further, a cause (e.g. smoke) can be conceptualized in different ways, which may lead to different lexical choices, as will be elaborated in the next section. Thirdly, the conceptual independence of the caused process is reinforced by the (prototypical) temporal distance between the instigation and the consequences (cf. also our comments on starve). Moreover, as is also typical of an ergative conception of events, there is a low salience of goal-achievement, which contrasts sharply with the foregrounding of this property in the case of prototypically transitive verbs, like the MURDER verbs but also strangle and throttle. The latter two verbs almost invariably code a process which ends in the death of the victim, whereas the ergative SUFFOCATE verbs commonly do not. Consider, for instance, the common usage of choke in reference to (mostly nonlethal) swallowing the wrong way. The SUFFOCATE verbs also often occur in hyperbolical uses, in which speakers (deliberately) exclude from the conceptualization the end point of the process, focusing instead on the suffering of the Medium. In Chapter 4, section 4.3.1, it was pointed out, that such usage is also common with starve, when it refers to the state of being very hungry and excludes the lethal outcome from the conceptualization which for some speakers may be the default case. In other words, an ergative conception can come to focus strongly on the Medium's activity, thereby not only excluding the instigation but also the end-point of the process. Fourthly, the distinction between the external and internal parts of the respiratory system that are affected is an additional experiential factor which motivates the transitive/ergative distinction. It is quite logical that the external perspective lines up with the focus on the Agent, since, most naturally, the affected parts are those parts that can be easily accessed by an external agent. Moreover, the more external parts of the respiratory system are also those that are not really active in the respiration process. As is typical of transitives, smother, strangle and throttle altogether omit the notion of the victim's participation. As will be substantiated below, these verbs conceptualize the interruption of normal respiration as externally inflicted on a fully inert patient, although smother continues in one particular metaphorical use the earlier


153

ergativization. It is worth noting that when typically ergative verbs, like suffocate or choke, occur in contexts where the cause is clearly and exclusively external, almost invariably an effective construction is used. This is corifirmed by both diachronic and synchronic data. Although in need of further support and clarification, a strong indication of a correlation between the ontological characteristics of the situation and the lexical and constructional properties of the SUFFOCATE verbs seems sufficiently established. However, the neat distinctions as discussed here are not absolute. The diachronic study in Chapter 4, section 4.3.2 has already revealed the constructional flexibility of the SUFFOCATE verbs in the course of time. This flexibility also shows up in the evolution of the lexical coverage of the verbs, as will be substantiated in the next section. The starting point of the description of the lexical convergence is, ironically perhaps, the different etymologies of the verbs. However, these prove important to explain the renewed differentiation that characterizes the evolution of the SUFFOCATE verbs to present-day English. Section 6.3 discusses how the SUFFOCATE verbs have stabilized (at least temporarily) into lexically and constructionally different prototypes, a process which has taken some verbs to the periphery of the field or even out of it altogether. As a complement to the discussion of the MURDER verbs in the previous chapter, the description is organized in subsections focusing on a particular verb or pairs of verbs showing considerable semantic proximity (e.g. strangle and throttle). As a logical extension to this, the final section in this chapter will consider a specific paradigmatic change affecting the ergative verbs choke and drown, viz. a transitivization triggered by particles. 6.2 Lexical and constructional convergence in Early Modern English Although not always exactly traceable, the etymological origins of the SUFFOCATE verbs can safely be said to be quite different. The following paragraph offers a brief sketch of these etymologies, mentioning between brackets the year of the oldest attestation listed in the OED (note that for the Old English forms this is slightly artificial, as they are continuations of indigenous forms).53 In subsequent descriptions, some of these etymological features will be picked up again, as they are the source of some of the present-day prototypes. Let us begin with the two less frequent verbs suffocate (1460) and asphyxiate (1706). The former is semantically parallel to its Latin root suffocatus, the past participle of suffocare meaning "to suffocate". Asphyxiate, a more 'learned' word (significantly, it is the 'youngest' of the group), is an English verbal derivation of asphyxia, which in turn is derived (via Latin) from 53

In addition to the OED, the etymological descriptions are drawn from Partridge (1979) and Shipley (1945).

154

LEXICAL PERSPECTIVES ON TRANSITIVITY AND ERGATTVITY

the Greek word asphyxia meaning "pulse". Accordingly, it starts off with a specialized reference to a deficiency of oxygen in the blood. The verb stifle (1400) is of obscure origin, but probably related to Old French estouffer meaning "to suffocate", and in its early uses emphasizes difficulty of breathing. Also of obscure origin is choke (1200), probably related to cheek or Old Norse kjâlki meaning "jaw-bone". In any case, the concept 'intake of food' has been a salient component of the verb's semantics. The oldest OED citations of the item precisely refer to food that accidentally "fallyth in to the pypes of the lounge". Quite interesting is the early usage of Anglo-Saxon smother (1200), especially in its close affinity with smore and smorder or smo(u)lder, which all have a salient reference to smoke. The etymological origin of drown (1300) is unclear—the verb is probably an indigenous Anglo-Saxon word—yet the verb has always maintained a strong reference to submergence in liquid leading to death. Strangle (1303) came into English probably via Latin strangulare (meaning "to strangle") in its turn derived from Greek strangale, meaning "a rope or cord, a halter". According to the OED, the verb throttle (1400) derives from the noun throat via the verb-forming suffix -le. This may explain why, especially in the older OED attestations, throttle means "to kill by cutting or stabbing the throat". There are also a number of citations where dogs or hounds are said to have throttled an animal hunted down. These uses may also explain the frequentative suffix -le, specifying that the victim is shaken back and forth by the throat.54 In short, by the 15th century there was a collection of verbs which all included in their conceptual base the notion of suffocation. The differences between these verbs were subtle and quite early a great deal of interchangeability can be noted, for both the 'indigenous' forms and the loan words. In other words, the different verbs have been competing for conceptual space ever since their introduction into English. Note that in some cases, the competition has led to the demise of certain items. For instance, the Old English verb querken, derived from Old Norse kverca "throat" and related to Dutch wurgen and German würgen, meaning "to choke, to suffocate" no longer exists. For reasons of space, it is not possible to review an elaborate array of uses for each of the SUFFOCATE verbs (some subtle differences relevant to the present-day situation will be discussed in the following section). Instead, it may be revealing to consider, as indicative of the semantic proximity of the SUFFOCATE verbs, a special type of construction which was quite common in Middle and Early Modern English, viz. doublets, the pairing of two words to 54

Shipley (1945) suggests that the -le at issue is the instrumental or agentive suffix as in handle or girdle. Although meticulous analysis of historical data may shed more light on the issue, I think that Shipley's explanation is incorrect.


155

express the same idea. Here is an edited selection of some of the numerous examples: (6) (7) (8) (9) (10) (11) (12) (13) (14) (15)

Now shal he me strangle and cheke ... (OED, 1303) [if it] be put into wine and choked or strangled ... (OED, 1574) ... water quencheth & cheketh & strangleth feble fyre (OED, 1530) [Richard Hun] ... strangeled and smodered [criminals]. (OED, 1548) ... pillowes vnto their mouths, [they] smored and stifled ... (OED, 1548). He was stifled or smoldered betwene twoo fetherbeddes (OED, 1548) He that stifles and smothers his own prophecy (OED, 1687) ... the smoke may stifle and choke them ... (OED, 1601) ... doth the Sun his rayes ... Smother or choke? (OED, 1647) A rauing cloude ... threatnes...To smore and drowne him. (OED, 1585)

The reversibility of the conjunction in almost identical contexts, as with example (10) and (11), makes it difficult to pinpoint any semantic difference between the verbs. Note that these doublets are quite different from V and V combinations discussed earlier, e.g. shot and killed. On the use of doublets, Brook (1965) observes that it was a common tool for literary expression (next to, for instance, alliteration and rhythm) in Old English (homiletic) prose. She points out that this tradition may have been reinforced in Middle English writing when the Romance element in English vocabulary was establishing itself. Initially, the pairing of the loanword and its native equivalent may have been of the glossatory type, "as if the author were glossing the loanword for the benefit of such of his readers as might be unfamiliar with it" (Brook 1965:66). As the loanword became more widely known, the need for such pairing lessened, yet the tradition persisted. Note, however, that some examples pair two indigenous words (e.g. smother/choke· or smore/drown). Even when one of the two is a loanword, there is no real suggestion that the one is a gloss of the other (given once again, the reversibility). In some of these doublets, the original meaning of a verb may be an extended one. For example, both choke and smother/smore/smolder occur next to strangle (external constriction) or next to drown (immersion). These extensions are not exclusive to these doublets, however, since they also occur independently. This further indicates the near-synonymy or at least the semantic interchangeability of these verbs. The high degree of lexical convergence is probably due to the ontological influences as explained above. Despite some overt differences, like submersion

156


versus constriction, the causes of suffocation are not always clear. Also, in the real world, the caused process of suffocation mostly stands out prominently, regardless the cause. In other words, it is this process which is experienced in almost all cases. As explained Chapter 4, co-occurring with the lexical convergence is a constructional one, as the ergative model seems to be activated for all verbs. While ergative usages of strangle or throttle may seem incredible to the modern ear, their occurrence in an earlier stage of English should not be seen in isolation. Let us, by way of example, briefly consider the ergativization of strangle as a case where lexical extension precedes the constructional one. For example, despite an early (maybe idiosyncratic) ergative use of strangle, (I praye God, if it were so, I strangle of Pis brede, 1338), the OED data indicate that the verb showed considerable lexical flexibility before the emergence of ergative constructions (from the 17th to the early 19th century). These lexical extensions involve a relaxing of the cause specification: in addition to suffocation resulting from constriction of the throat by a rope-like instrument, the verb came to code suffocation resulting from various types of causes, such as poisonous gases, swallowing the wrong way, hanging, or drowning, as already illustrated in some of the doublet-examples above. However, quite early also, the verb came to refer to more general killings (i.e. non-suffocation events), such as deaths caused by poisoning (another hardly perceptible cause), by the sword (rare), or by violent assaults by wild animals or devils: (16) (17) (18)

Hanybal... stranglyd with poisoun (OED, 1443) The swearde shal strangle them (OED, 1535) The Jew ... strangled was of a leon (OED, 1390)

The more general uses of strangle were well established by the beginning of the 16th century. So, it appears that these extensions, especially those bringing in other causes of suffocation, have paved the way for the subsequent ergativization of strangle which almost invariably occurred with internal causes, a process through which the verb lined up with verbs such as suffocate, choke or drown. An interesting case of ergativization complicated by lexical competition is that of smother and the related verbs smore and smolder. Of these, smore seems to be the oldest (from OE smorian); in early ME the variant form smother appears. The Middle English Dictionary postulates that smo(u)lder is probably derived from the noun smorðer (meaning "smoke, smoky vapour") via r/l dissimilation. Although the notions of smoke, heat or fire remain salient, all three verbs show considerable semantic proximity to other SUFFOCATE verbs such as stifle, choke, drown and strangle (especially in its extended usage). Plausibly, this can be accounted for by the fact that smoke is commonly


157

conceptualized in different ways. It can be seen as a gas polluting the air and lethally entering the lungs (prototypically the conception underlying suffocate, choke and stifle). Alternatively, it can be conceived as cutting off fresh air, which renders a more external point of view (moving closer to the conceptual profile of strangle). Or, it can be conceived as an engulfing wave (a conception prominent with drown). A particular extension common to all three verbs, which has left constructional traces as well, is the reference to a fire burning or made to burn low, a natural extension, given the experiential proximity of (thick) smoke and a low burning fire.55 Like in the other uses, the ergative pattern also shows up here, as illustrated by the following examples: (19) (20)

Set on fire a quantitie of haye ...and whiles it is smoothering and smoaking ... (OED, 1667) This woode burnetii nat clere, it dothe but smolder. (OED, 1530)

Also in this usage there is considerable influence from the ontological situation: a fire or an entity on fire is readily conceived as a semi-autonomously acting entity, i.e. burning low by itself, without necessarily implying an Instigator. The ergative conceptualization thus lines up nicely with our experience of the world. The brief diachronic survey has revealed that the SUFFOCATE verbs displayed a great deal of lexical interchangeability and constructional convergence into a general ergative pattern. In some case (e.g. strangle) lexical extensions could be regarded as paving the way for a later ergativization. However, separating the lexical and constructional dimension and asking which extension came first, seems to be asking the wrong question (much like the chicken and the egg). Rather, my view is that the lexical and constructional meaning are interdependent. In most cases, semantic modulation is due to the dynamic interplay of lexical and constructional meaning, in which it is no longer clear what triggered what. In any case, the high degree of lexical and constructional flexibility of the SUFFOCATE verbs has decreased somewhat in their further evolution to the present-day situation, which can roughly be characterized as a process of reorientation towards their etymologically basic meanings. The lexical and constructional differentiation will be taken up in turn in the next section. 55

Example (8), water quencheth & cheketh & strangleth feble fyre, shows that other SUFFOCATE verbs could also express the meaning of extinguishing a fire. Of these three, only quench (from Old English *acwencen "to extinguish") has retained this meaning, although nowadays it too is used predominantly metaphorically, usually in the meaning of "to satisfy one's thirst, desires and the like" (Collins Cobuild). The further evolution of smoulder will be discussed shortly.

158

LEXICAL PERSPECTIVES ON TRANSITIVITY AND ERGATIVTTY

6.3 Lexical and constructional differentiation in Contemporary English 6.3.1 Overview of the SUFFOCATE verbs Before embarking on a more systematic treatment of lexical and constructional differences between the SUFFOCATE verbs in present-day English, it may be useful to have a look at the frequency distribution of these verbs in the contemporary corpora (WSJ, LDC, and CASS), presented in Table 8: LEXICAL ITEM

drown choke strangle smother

stifle suffocate throttle asphyxiate TOTAL

FREQ

PCT

144 141 53 46 47 30 27 5 493

29.2% 28.6% 10.8% 9.3% 9.5% 6.1% 5.5% 1.0% 100.0%

Table 8: Frequency of the SUFFOCATE verbs in the contemporary corpora It should be recalled that subsumed under the heading "lexical item" are all verb forms, nouns (e.g. strangler, asphyxia) and compounds (e.g. stranglehold)', the latter two are only a minority. The frequency of the items as well as that of their different meanings is important to present a correct picture of the data, as will be illustrated, for example, in the discussion of strangle and throttle below. It should be noted that contemporary attestations from the OED are not included in Table 8. 6.3.2 Suffocate Starting with the general verb suffocate, it can be noted that the verb has remained relatively stable in its usage continuing to profile the physiological process that in fact is part of the base of all SUFFOCATE verbs, which renders it all the more appropriate as the name for the entire category. From a pure lexical-semantic point of view, one could characterize the relationship between suffocate and the other members as metonymical, because the verb profiles only a part of the complex conceptions designated by the other verbs. The salience of the physiological process in the Affected may explain the preference for a non-effective construction. While an effective construction with a human Instigator is possible, historical and contemporary material are characterized by its unmistakable absence, even in contexts where a human participant can be readily traced as having ultimately caused the suffocation, as in the following sentence:

THE MEDIUM-CENTREDNESS OF THE ERGATIVE PARADIGM (21)

159

He pushed her face in a lemon meringue pie and held it there until she suffocated (CASS)

An effective construction seems to be restricted to situations with an Instigator of circumstantial nature. These Instigators are often situations or gaseous substances that cut off fresh air, e.g. a heavy, colorless gas that in large quantities can suffocate humans (WSJ). Despite being conceptually basic within the SUFFOCATE set, the verb suffocate has a low frequency, in all the corpora consulted. Why this is so, is difficult to explain. Unlike the other SUFFOCATE verbs, suffocate is manner or means independent and highlights primarily the activity of the Medium. Its generality explains why it serves as the cover term for the entire category yet perhaps also why it is so infrequent (with a marked preference for literal usage). The reason for this is that there may be a tendency to use either a neutral coding (e.g. in terms of kill or die) or a specific one (e.g. strangle or poison), rather than a coding that holds the middle between the two as suffocate does. Additional research is, however, warranted to test this hypothesis. 6.3.3 Asphyxiate We can be brief about asphyxiate which continues its specialized reference to a deficiency of oxygen in the blood. As with suffocate, this conceptual structure (for which a coding with the noun asphyxia or asphyxiation is in fact most common) can be seen as a subcomponent of the other SUFFOCATE verbs, and indeed is often used as such: (22) (23)

When a man is strangled, drowned, or choked ... what is called asphyxia comes on (OED, 1872) He may have been smothered leading to asphyxiation. (sw)

However, both noun and verb have been extended to a more general use virtually synonymous with suffocate. The OED does not mention the possibility of a non-effective construction for asphyxiate, yet in present-day English, a non-effective construction does occur, possibly resulting from this more general use. Consider the following examples: (24) (25)

... we will shoot out into deep space I expect and asphyxiate. (HHGG) ... a ship whose two-man crew was at that moment lying asphyxiated in a smoke-filled chamber several miles beneath the ground. (HHGG)

As is typical of past participle constructions with ergative verbs, the second example is ambiguous between an effective and a non-effective reading.

160


Comparably, a drowned man can mean "a man who drowned" or "a man who was drowned". In the overall description of the SUFFOCATE verbs, asphyxiate plays a negligible role, since it can be taken to be a more technical variant of suffocate, restricted to medical prose or a more formal register. 6.3.4 Smother, smoulder, and smore Concerning the closely related verbs smother, smoulder, and smore, the most drastic change is clearly the demise of smore which, according to the OED, now only occurs in Northern dialects. The gap left by smore was happily filled by smother. Smoulder emerges as strictly intransitive and is exclusively used in reference to a low burning fire or metaphorical extensions thereof. Of the three, smother is the only verb that in present-day English can still refer to actual suffocation, but it too has changed, centring less around the notion of smoke and emphasizing more the notion of coverage. Two important factors can be suggested to account for this conceptual shift. First, smoke especially when thick is often conceptualized as a covering substance; further evolution to the present-day situation has come to extend the conceptualization of coverage at the expense of the original prototype. Secondly, probably triggered by the same mechanism, later uses of smore had more saliently taken on the notion of coverage as well, as for instance reflected in the verb's usage in reference to cooking in a close vessel, e.g. on a gentle fire ... let it stew, and smoar till the hearbs and the onyons be soft, OED 1615). Some hundred years later, also smother was used in this meaning. From the conceptual competition between the two verbs, smore came out the loser (at least formally) being ousted from standard English; smother emerged with a new prototype resulting from a "conceptual merger" of the two verbs (cf. Geeraerts 1988), leading to an increased prominence of the notion of coverage. Here are some typical examples of present-day smother: (26) (27)

He seized the blanket from the ground and advanced towards Aanoch, poised to smother the bound man. (CASS) ... I was just so relieved that Ms. Gruberova ... didn't accidentally smother herself in a drape. (WSJ)

Some extensions of smother go even further and drop the negative connotation altogether, retaining only the conception of an entity being (fully) covered, or at least its visible parts. A typical example is the following sentence, part of a nostalgic description of Wrigley Field, Chicago: (28)

The verdant ivy that smothers the gently contoured, brick outfield walls still swallows bouncing drives. (WSJ)


161

If there are negative connotations, they draw on the excess of the coverage, e.g. a smother crop to kill out weeds (OED on CD-ROM). In line with the external character of coverage, smother behaves like a fullfledged transitive verb in present-day English. However, it has been attested in an ergative construction when used in a metaphorical usage mapping the image of a smouldering fire onto the emotional domain, a usage which had been quite common with this verb, e.g. Heere began the harme to smother like wet hay in fire (OED, 1621). In a series of contemporary short stories by the same author, the following instances of smother were found: (29) (30) (31)

The sharp jab of cold steel replaced the smothering pain... (CASS) [the loneliness] swept upwards, ... knotting itself around his soul and slowly squeezing until he could feel the suffocating, smothering, nothing. ... the ground came crashing back upward, spinning as it impacted and smothered.

The examples reveal a mixed pattern: examples (29) and (30) realize an ergative non-effective, emphasizing the slowness of the coded process (cf. the contrast between sharp jab and smothering pain). The construction in (31) is an objectless transitive, construing the ground as Agent. However, given that the non-effective has not been attested in any of the other corpora, this may be an idiosyncratic use, and in Figure 18 above, the verb is categorized as a transitive, with the ergative potential between brackets. Further analysis of a large collection of contemporary data is warranted to determine the salience of ergative constructions such as in (29) and (30), which may occur more frequently when smother comes to code events for which the notion of coverage is less salient (thus encroaching on the semantic space of other SUFFOCATE verbs like choke, stifle or suffocate). In any case, it is interesting to see how for a particular speaker there appears to be a paradigmatic split determined by the verb's meaning. 6.3.5 Drown In some cases, smother comes close to drown when it implies total containment in a liquid as in the following cases: (32) (33) (34)

...who likes his bacon sandwiches ... smothered in ketchup (LDC) Ahab half smothered in the foam of the whale's insolent tail (NOV) she was drowning ... thick, crude, black oil smothering her ... (CASS)

162


Although total immersion is implied here, enforced by the containment expressed by in for the first two examples, the uses cannot be fully equated with prototypical drowning, as the substance is not truly liquid, it is either relatively thick (ketchup and oil) or non-liquid at all (foam). The conceptual spaces smother and drown overlap in what could be called a grey transition area, i.e. in uses that extend from the prototype. As said, smother comes close to drown in the case of coverage with barely fluid liquids, as in the examples above. Drown comes close to smother in the case of immersion in covering liquids, example (35) serving as the counterpart to (32) and (34), or in the case of metaphorical extensions, as in example (36): (35) (36)

... a roast beef au jus wave ... drowning him (CASS) a division ... was drowning in paper work. (WSJ)

Strikingly, drown retains a strong link to the image of immersion even in metaphorical extensions, as some typical examples show: (37) (38)

...MiniScribe was drowning in red ink. (WSJ) ... she just drowns you in verbal diarrhoea. (LDC)

The overlapping between smother and drown is quite natural given that both appeal to the concept of coverage, albeit to a different degree (where immersion could be regarded as full coverage), yet the common ground is situated in uses that are atypical for both verbs. In connection with the ergativity of drown, it can be noted that the data from the contemporary corpora refute the claim made in the OED that the noneffective is "now unusual": more than half of the attestations for this verb are non-effective. The semantic content of drown almost predicts its ergative construal potential: the conceptualization of a drowning is grounded in our experiential knowledge that such an act is not necessarily being caused by an 'active' cause. Typically, the verb's profile is restricted to the drowning entity which does the actual work. Significantly, even in effective construals, the Instigator is often only vaguely defined, as in the following examples: (39) (40) (41)

...the story of Narcissus, who ... plunged into [the fountain] and was drowned. (NOV) Hannah's husband was drowned whilst swimming his horse-team across the flooded river. (OED on CD-ROM, 1908) About 11,000 sea turtles are drowned every year in U.S. waters as a result of entrapment in shrimping trawl nets (WSJ)

The passive construction clearly indicates that some Instigator was involved, but this Instigator remains fairly vague and is in fact no more than a hint. These


163

agentless passives with drown contrast sharply with those with the MURDER verbs (e.g. Many people were murdered/assassinated/executed) which even then prototypically evoke the conception of a human, volitional Actor. As said in Chapter 5, Davidse has observed that the ergative paradigm does not have the fundamental hook-up with an intentional Agent that typifies the transitive paradigm. In an ergative construal of an event, unintentional Agents are quite common and, as said, often the Instigator remains quite vague. An interesting illustration of the presence versus absence of a strong commitment to an intentional Agent is furnished by strangle. In contemporary English, the verb is strongly transitive; hence, an Agentless passive as in she was strangled still saliently implies a volitional Agent. In our earlier diachronic description, it was shown that in an earlier period, strangle showed considerable lexical flexibility and allowed the ergative pattern. Strikingly, in this period Agentless passives as in (42)

A Labourer, who was sent down [the well] to recover it [a hammer]; ere he reach'd the Water, was strangled. (OED, 1728)

did not really evoke the notion of an Agent either. Inquiry into the Instigator seems ill-founded, much like with drown in examples (39) and (41). The situation can perhaps be compared to atypically transitive kill in agentless passives in the context of accidents, e.g. many people were killed in the crash, for which, it will be recalled, the inquiry as to an Agent seems ill-founded. However, the data for drown suggest that such an agentless passive was common in earlier stages of English (up until the 19th century, as indicated by numerous attestations for that period in the OED and the NOVEL corpus), but that in contemporary English it is unusual, example (41) being an exception. In literal use, a non-effective is most frequent as is an ambiguous past participle construction (e.g. a drowned corpse), although less so. Effective constructions do occur, sometimes expressing the volitional human Instigator responsible for the event, as in (43)

Suppose he should take it into his head to duck you, ... yes, and drown you (NOV)

Nevertheless, effective constructions mostly occur with an Instigator whose enveloping characteristics are somewhat unusual. They are mostly conceptualized as moving objects, waves, or the like, as in (44) (45)

Failing to drown me, the wall of water ... (CASS) ... a roast beef au jus wave ... drowning him (CASS)

164


It can be reiterated, however, that such effective constructions are not so frequent and occur in contexts where drown comes close to the semantic space of smother. Most typically, a non-effective construction is used, with the enveloping substance sometimes surfacing as the setting within which the event Takes place. (See, however, further comments on the drown out construction in section 6.4 below). 6.3.6 Stifle Another remarkable evolution within the field concerns the verb stifle. As said, the etymological origin of the verb emphasizes difficulty of breathing, which may eventually result in the victim's death. The difficulty arises from the lack of fresh air, mostly caused by the 'pollution' of the air through smoke or heat, by being closed in, or because the nose and mouth are covered. Since these features also figure prominently in the conceptual structure of smother (and to a lesser extent in that of choke), there is considerable overlapping, as emerging from the OED data which also show that doublets with these two verbs were quite common.. However, careful analysis of the OED data indicates a slight difference between stifle and smother that will trigger the later semantic divergence: stifle more readily occurs with 'coverings' of the atmospheric type. More often than with smother, it is the location or the quality of the air within it that is seen as the cause of the stifling. Gradually, stifle and smother will grow apart to different centres: the latter will, as argued above, come to emphasize the physical coverage whereas stifle will further crystallize the notion of 'atmospheric pollution' to a central meaning component and usually exclude the lethal outcome altogether. The latter has been common with stifle throughout its history. One early attestation of stifle, for instance, says that people were stuffled because of a great heat, so that they felle into a grete sikenesse, yet they apparently did not die (at least not immediately). Hence also the commonality of hyperbolical or exaggerated use. The contemporary fiction corpora further substantiate this semantic specialization of stifle: (46) (47) (48)

The heat is stifling. (LDC) The sudden closeness of the bodies of all his schoolmates made the air so stifling that Johnny was not able to breathe. (CASS) The [boys] ... were cruelly consigned to a stifling classroom. (CASS)

In all cases, the atmosphere is causing the stifling. The all-enveloping property of the atmosphere further distinguishes stifle from smother. This atmosphere (or some characteristic of it or its location) is usually (but not necessarily) cast in


165

the Agent role, as demonstrated by heat, air, and classroom in examples (46) to (48) above. A chronological ordering of the citations from the OED on CD-ROM reveals that stifle has also taken on a marked preference for metaphorical usage. For instance, of the 30 attestations from the 19th century, only one is a literal usage. The contemporary preference for metaphorical usages is reflected in the corpora: in the WSJ corpus all 20 attestations are metaphorical; in the short stories, (47) is the only literal usage out of 14 attestations. Most uses have an abstract entity in the Agent role, often a reified conception of an event or action. Here are some representative examples: (49) (50)

... the move will "stifle innovative changes" and slow medical progress (WSJ) Many things, from oil shortages to trade wars, can serve to stifle economic growth. (WSJ)

So, from a lexical semantic perspective, it can be argued that stifle is in fact moving out of the domain of verbs of killing, which is probably why Levin (1993) marked the verb with a question mark. As a relic from its earlier use, contemporary usage may still highlight the difficulty of breathing which produces a choking sensation but is not lethal. As to its constructional characteristics, it will be recalled from Chapter 4, section 4.3.2, that the OED lists an ergative non-effective as late as 1966: The garrison stifles in rock heat. It is to be doubted whether this is still a productive pattern; hence, in Figure 18 the verb has been classified under the transitives yet closer to the ergative side, given the latent possibility for an ergative construal. 6.3.7 Strangle and throttle Two other verbs whose historical development shows considerable overlapping are strangle and throttle. In fact, even in present-day English the two verbs appear virtually synonymous, as witnessed by the following examples: (51) (52)

"I'll throttle him with his entrails!" the Advisor declared. (CASS) May the last king be strangled with the guts of the last priest (LDC)

(53)

... NCNB is busy throttling ... customers with the other hand. (WSJ)

(54)

She raised her hands as if to strangle the chunk of dirty ice (CASS)

Despite the near-synonymy in these uses, some subtle differences between the two verbs may be discerned which in fact continue the different etymologies. First, throttle seems to be less tied to circular constriction, a feature quite

166


prominent for strangle, retained from its original reference to strangulation by a rope. Compare: (55) (56)

The soldier then drew a rifle across his neck and pulled back, throttling him as [he] worked on breaking the lieutenant commander' s neck (CASS). *?The soldier then drew a rifle across his neck and pulled back, strangling him as he worked on breaking the lieutenant commander's neck.

The construction with strangle is odd, since it has the strong implication that the rifle was bent around the victim's neck. A second subtle (yet non-rigid) semantic difference lies in the metonymical ambiguity that the verb throttle has inherited from its nominal root throat: the focal area may be the outside, the anterior part of the neck (compare he grabbed me by the throat) or the inside, the pharynx (compare I have a sore throat). This metonymy has guided the evolution of throttle up to the present, to the extent that in literal usage an outside point of view is most common, whereas in metaphorical usage the internal perspective seems more prominent. The latter is apparent in the metaphorical extension to the suppression of the voice (or more generally, sounds emitted from the throat), an issue to which we will return in the conclusion of this section, as it is common to all the SUFFOCATE verbs. The subtle differences between strangle and throttle, explained on the basis of their etymology, are summarized in Figure 19.

Fig. 19: Profile and active zone of strangle (a) and throttle (b) The grey area marks the active zone within the profile indicated in bold. For strangle, the rope-like character of the constricting element is represented as a flattened circle; the rectangular with throttle stands for any type of constrictor. I hurry to add that the figure should not be interpreted as implying a stable and rigidly delineated dichotomy; often, the active zone may not be more than a hint. As already pointed out, the two items have been encroaching on each other's conceptual space ever since their introduction into English. Nevertheless, the data in the corpora suggest that the items still centre around these different foci.


167

A third element in which the two verbs differ is their degree of goalachievement: in contrast to strangle, goal-achievement is less saliently for throttle. The following example confirms the point: (57)

I don't want to kill you, Mike. I want to throttle you, and then I want to kill you. (CASS)

In this view, it becomes clear why throttle often occurs in contexts where the action of throttling is done "for the purpose of subduing or overcoming resistance" (Webster's 1978). Suppression of resistance as well as non-fatality also figure in the following example: (58)

China's largest city remains uncowed by the throttling of the country's pro-democracy movement 10 weeks ago. (WSJ)

This sentence, the opening sentence from an article about Shanghai, states that despite the bloody suppression of the protest movement in Beijing, the city of Shanghai remains uncowed, i.e. not yet frightened into submission. The following sentence in the article substantiates this claim: Unlike Beijing's subdued populace, many Shanghaines openly express resentment against the government and wonder where their lingering spirit of rebellion will lead them, (my emph.) This seems to suggest that throttle is a less prototypical transitive verb, given the prototypicality of goal-achievement within the transitive paradigm. Despite this difference in degree, both verbs are full-fledged transitive verbs, as the earlier ergativization has not entrenched. An exception is the ergativization of throttle down an engine which, however, was shown to be independent of the global ergativization of the SUFFOCATE verbs in (Early) Modern English. Finally, an important difference between the two verbs is their distribution and frequency. Both historical and contemporary data show that strangle is far more frequent than throttle. Striking also is that throttle has apparently never been attested in a doublet, which signals its low degree of entrenchment. Further, strangle is far more outstretched in its semantic coverage, especially when the historical data are taken into account (cf. above). The OED not only distinguishes many more different meanings for strangle than it does for throttle, it also furnishes manifestly more illustrations for each of the meanings distinguished. Strikingly also, as agentive noun it is strangler and not throttler which has entrenched as the term to refer to a (serial) killer. This once again signals the non-prototypical status of throttle. In contemporary English, this appears to be reinforced, at least for some speakers, by the verb shifting its prototype to the meaning "control the fuel in an engine", taking the verb's conceptual structure outside the field of killing

168


altogether. Note that this may also entail a reinterpretation of the metaphorical uses: throttling a protest movement is then interpreted as a metaphorical mapping from engines to abstract entities that are conceptualized as moving objects whose movement can be controlled. How these speakers see a usage such as I'll throttle him with his entrails referring to caused suffocation remains unclear. Further (psycho-)linguistic research may shed light on whether this is perceived as a semantic extension from the prototype or as a usage homonymic with it. In brief, careful analysis of historical and contemporary data reveals that two virtually synonymous verbs, strangle and throttle, can be seen to imply subtle conceptual differences. 6.3.8 Choke The last of the SUFFOCATE verbs whose rich semantic history deserves some attention is choke. Its historical expansion is impressive indeed, spanning almost the entire field covered by the SUFFOCATE verbs in seemingly equally frequent and well-entrenched extensions. Strikingly, throughout its history, its etymologically basic reference to swallowing the wrong way has remained quite salient. Also in present day English, this usage is prototypical. In fact, from a diachronic perspective, choke provides a fine illustration of what Geeraerts (1994:138) has called "semantic change starting from subsets", where it is not the category as a whole but a salient subset within that category—in this case, the concept of food getting stuck in the windpipe—which plays a pivotal role in semantic change. Consider some typical usages of this verb: (59) (60) (61) (62) (63)

Riddle laughs so hard he starts to choke on his salad. (CASS) Sorrille almost choked on his tongue (CASS) "... it was supposed to mean they'd be LUCKY together." He choked on lucky. (CASS) ... a Third World slum choking on its own refuse. (WSJ) "He is choking on his own debt," the executive said. (WSJ)

Example (59) instantiates the verb's prototypical use. Although prototypically the choking is caused by food, example (60) illustrates it can be another unbreathable substance. Example (61) is a common metaphorical usage and envisages words as entities whirling in the throat, and potentially clogging it (this metaphor will be discussed shortly). More overtly metaphorical are (62), in combination with a concrete object (refuse), and (63), in combination with an abstract object (debt). When used in the prototypical meaning of food blocking the windpipe, or in metaphorical extensions thereof, the verb invariably occurs in an ergative noneffective construction, with typically the cause specified in an on-complement.


169

This is an interesting phenomenon: why is a typically overt cause coded with on and not made explicit as the Instigator in an effective construction? In fact, when used in this meaning, an effective construction is not acceptable: (64) (65)

*the salad choked Riddle. *Sorriile was almost choked by his tongue.

Secondly, why should English settle for the preposition on when the object is generally known to actually enter into the throat and windpipe? The easiest way out would be to stipulate that the use of preposition on is simply an arbitrary convention of the English grammar, deviant from other languages. Modern Dutch, for instance, appears less arbitrary and harmonizes more with the ontological situation: the preposition in (equivalent to English in) is to be used in comparable expressions, e.g. zich verslikken in een aardappel ('to choke on a potato"). The same sentence appears in Spanish and Portuguese with the preposition con/com (Spanish: atragantarse con una patata; Portuguese: engazgar-se com a batata). Like its English equivalent with, this preposition prototypically expresses the instrument. (The preposition with can also occur with English choke, a construction to which I will turn shortly.) German takes yet another preposition, sich an einer Kartoffel verslucken; the preposition an is not unlike English on, but still differs from it. However, the fact that different languages express the unbreathable object differently does not warrant the conclusion that the choice of prepositions is arbitrary. The crosslinguistic variation of these prepositions is a good illustration of how different languages incorporate different conceptualizations of the same reality. Undeniably, the use of the preposition on in English is a language-specific convention, yet one that is conceptually well-motivated. My claim that the use of on is well-motivated is inspired by Herskovits's (1982, 1988) analysis of some spatial prepositions in English. She analyses on as ideally designating a relationship of contiguity and support between two entities56: for a geometrical construct χ to be contiguous with a line or surface γ; if γ is the surface of an object Oy, and χ is the space occupied by another object o x , for Oy to support ox. (1982:186).

She points out that on may, however, extend to be used in a context where containment is at issue, and where one would expect the preposition in to be 56

Herskovits's notion of ideal meaning refers to the conceptual centre around which a semantic category is organized. It is thus equivalent to our notion of prototype, a term she finds too laden and possibly inviting misunderstanding.

170


used, as is illustrated by the alternations the dust in/on the howl or the oil in/on the pan. Which option one chooses, Herskovits reasons, is determined by what is deemed most relevant, the idea of containment or that of contact. Herskovits5 s notion of relevance provides the motivation for on to appear with choke because in this context the notion of contact between the eater's mouth/throat and the substance is of immediate relevance. In this extension, the support condition is dropped from the preposition's ideal meaning and the notion of contact is maximized. The use of on in choke on is comparable to its use in to blow on a whistle, to suck on a vitamin tablet, or chew on a chewing gum. It is impossible to conclude from the OED data whether the use of the oncomplement actually originated with the intake of food only later to extend to constructions like choke on his {tongue/words}. In any case, I take the notion of forceful contact with the throat and the subsequent gagging as central to the use of on (which suggests that it is indeed original to the meaning of swallowing the wrong way). Further, for this meaning of choke, the salience of the internal contact and the subsequent gagging triggers a non-effective construction, maximizing the Medium's participation in the process. Interestingly, when both the salience of (forceful) internal contact and gagging are reduced, an effective construction becomes more typical. Both historical and contemporary language use confirms this hypothesis: an effective construal typically appears in contexts where there is no internal contact: (66) (67) (68) (69)

Slaine or some say Choaked by his own men (OED, 1584) The broider'd band That underbraced his helmet at the chin...Choaked him. (OED, 1790) The man who choked the Emir (OED, 1866) ... a solution for automobile pollution that is choking urban areas. (WSJ)

The examples in fact speak for themselves: people and helmet bands are by definition conceived as constricting the windpipe from the outside. The same applies to (69): it is the general condition of pollution that is affecting the suburban area. One should, however, not misinterpret these observations as indicating that the ontological situation rigidly predicts the construal. Nevertheless, the data unequivocally reveal a strong experiential motivation for the occurrence of an effective or a non-effective construction. Interestingly, the OED citations on the -er nominal choker confirm my point of view. All citations of the agentive nominal profile the Instigator, and almost invariably, it concerns an external Instigator. Despite the possibility of a truly external cause (e.g. the constriction of the throat, see example (71)), the contemporary data seem to indicate this is


171

uncommon; mostly, the causes are of the 'non-breathable' type that enter the throat and thus lead to breathing problems or suffocation.57 The bodily reaction to the entrance of an unbreathable substance into the trachea, i.e. that of retching and heaving, is quite prominent and is often singled out for profile in the context, as in the following examples: (70) (71)

Ginny chokes on her coffee. She rises, coughing ... (LDC) ... suddenly he'd ... grab him by the throat and choke him till he coughed up an additional $2 billion or so. (WSJ)

In (71), the image suddenly shifts from the literal to the metaphorical domain with the addition of an additional $2 billion or so. The salience of the bodily reaction has given rise to extended uses of choke which refer either to this specific bodily action or to the noise of such action, as in the following examples: (72) (73) (74)

Miss Squeers ... was ... taken with one or two chokes and catchings of breath. (OED, 1839) "Not important," she says again with a loud, choking hiccough. (CASS) Haezar made a choking noise. (CASS)

In contemporary English, choke frequently profiles the retching and backgrounds the suffocation, mostly leaving the latter out of the scope altogether. This may also explain why choke often occurs in the -ing form which imposes an internal perspective on the process by excluding its end points, thus downgrading the eventual suffocation. The forceful contact and subsequent gagging that is saliently incorporated into the semantic structure of choke (on) finds a striking counterpart in choke with, a pattern quite common in present-day English, especially in a past participle construction. Essential to this usage is the contribution of with, whose meaning can informally be identified as designating "addition", related to usages as coffee with cream or its usage in combination with 'SUPPLY' verbs, such as load, burden, fill, pollute or infest. In the case of choke, the addition is typically conceived from an internal perspective (comparable to a construction like my eyes filled with tears) or as coming in through the throat and gradually filling and/or clogging the windpipe from within. In contrast to the forceful and acute physical contact implied by choke on, the preposition with implies the notion of gradual saturation and excess. The implication manifests itself quite early, it 57

An exception is Scottish English where choke is the common term to denote external constriction and is thus the equivalent of Standard English strangle or throttle.

172


appears in the very first metaphorical attestation listed in the OED: so chokid wip talow of worldly goodis ... Þat Þei may not preche pe gospel (1380). Clearly, the substance added is harmful, yet its impact appears of the chronic type. Logically, the energetic bodily reaction of retching and heaving is backgrounded in a conceptualization with choke with. The difference between choke on and choke with can simplistically be characterized as follows: choke on

substance in

forceful blockage

choke with

substance in

gradual saturation

substance out (

substance out)

The contrast in the nature of the contact may also explain why the participial construction, designating the final state of the process, is possible for with (he is choked with emotions) but not for on (*the company is choked on debt). Additional support for my analysis of choke with is given by DeLancey who says about with as used in He is in bed with the flu that it "marks conditions which persist over time, and thus for a certain time are co-locational with the patient" (1984b:205). Thus, he notes, "one cannot say die with pneumonia because at that point the disease and the individual are no longer co-locational" (1984b:205,fn. 26). In the contemporary corpora, choke with almost invariably occurs in metaphorical readings, especially in reference to the emotional domain. Typically, the affected entity is conceived as a stream or pipe-like container being clogged or filled up by the addition of harmful substances. Consider some typical examples: (75) (76)

I nearly choked with longing for that ... (LDC). The toilet was choked with paper and cigarette butts. (CASS)

Once again, despite some prototypical patterns, there is no absolute predictability for a certain coding, as illustrated by the following contrast: (77). (78)

...the Ashless Vistula, the river choked by pollution. (WSJ) By the time it reaches West Germany, the river is choked with sewage, mercury and other chemical wastes. (WSJ)

In these cases, an ontologically similar situation is coded slightly differently: arguably, example (77) emphasizes the general condition of pollution as deadly affecting the river, whereas (78) can be said to emphasize the pollution as resulting from the gradual addition of chemical substances, as also indicated by the opening phrase by the time it reaches West Germany.


173

Although a non-effective construction for choke with is possible, as shown by example (75), it is not so common. Typically, a past participle construction is used, focusing on the outcome of the gradual choking process. A past participial construction without a preposition, as in (79) (80)

... [they] tried to fight their way through smoke-choked hallways ... (WSJ) ...a village of unpaved streets and garbage-choked canals. (WSJ)

has two possible paraphrases: "hallways/canals choked by smoke/garbage" or "hallways/canals choked with smoke/garbage". This ambiguity of past participle constructions reflects the voice syncretism that is typical the ergative noneffective. To round off the description of choke, a specific extension should be mentioned in which the verb, but more particularly the noun, is parallel to throttle, viz. its reference to controlling an engine. However, with choke it is not the amount of fuel that is regulated as is the case with throttle, but the amount of air, necessary for the combustion of the fuel. Insufficient influx of air will stop the engine (it will be flooded). Notice the non-effective usage in the following example. (81)

... I would have [the mower] choke or run out of gas. (CASS)

So, as to air and fuel control in engines, choke and throttle have become complementary, and both allow the engine to be conceptualized as selfinstigating the process. It will be recalled from the earlier description that for throttle this is a particular ergativization, whereas choke merely continues its prototypical ergative potential. 6.3.9 Summary The above discussion of the semantic coverage of the different SUFFOCATE verbs has illustrated that in their evolution to contemporary English, the verbs have crystallized towards more distinct prototypes, mostly a continuation of their etymological basis. This may have been accompanied by some conceptual reshuffling, as was illustrated for smother. The differentiating feature is mostly the type of cause that brings about the suffocation. Excluding the more general verb suffocate and its more formal equivalent asphyxiate, the prototypical properties of the SUFFOCATE verbs can be identified as follows:

174


choke drown stifle smother throttle strangle

internal obstruction of the windpipe/throat immersion in water lack of fresh air in atmosphere coverage of nose and mouth external constriction of the throat external, circular constriction of the throat

It be further noted that especially for stifle and choke a non-lethal outcome is implied, although for the latter, such may still be the case. However, the question arises whether by drawing up these distinctions in prototypes we are not trapped in seeing differences that may not be there. This brings us to an important observation made by Geeraerts, who warns against "the tendency [...] to look for purely cognitive or conceptual explanations of the facts that one encounters" and points out that "language is not just content: it is also form, and its formal side has an expressivity of its own" (1988:227). The existence of the near-synonymous SUFFOCATE verbs in the Middle English period such as smore, smolder, smother, stifle and suffocate could be taken to confirm Geeraerts' point, and the variation noted would then merely function "to express something conceptually old in a formally new way" (1988:227). Loanwords may often be used to enhance the rhetoric and not so much to express something semantically distinct from the indigenous equivalent. Yet also the near-synonymy οf strangle and throttle in present-day English could be seen as an instance of a merely formal difference. The overlapping is for instance also reflected in dictionary definitions which often use one verb in the definition of another and vice versa. Consider some of the definitions given in The New Lexicon Webster's Dictionary (1989): throttle choke

to choke to stop or almost stop from breathing by applying pressure on the windpipe or by blocking it, to asphyxiate strangle to kill by compressing the windpipe and so preventing breathing suffocate to kill by stopping respiration, to smother, to choke The problem is certainly an acute one for the frequent metaphorical extensions of the verbs, where the (ontologically-based) differences drawn above no longer apply. Consider, for example, a metaphorical extension that is common to practically all the SUFFOCATE verbs, viz. the suppression of the voice or sounds emitted from the throat. Here are some illustrations: (82) (83)

... with a choked cry [I] slammed my hooves into him He definitely heard a stifled cry from the creature ...


(84) (85) (86) (87)

175

... half-uttered screams that seemed smothered violently The howl turned into a strangled whine... ... a smooth basso, somewhat throttled ... Neville's cries were drowned out by the applause ...

The reason why this extension is common to most SUFFOCATE verbs is fairly obvious: in literal usage, the affected area comprises organs which are also part of the articulatory apparatus. The metaphorical extension to the suppression of speech appeals to the same physical area, but shifts the profile to the vocalization which is coded as the affected entity. One could thus argue that for these extensions metaphor is mediated by metonymy. Although some features of the source domain may be retained, the differences between the individual items are weakened. Another common metaphorical extension concerns the suppression of emotions: how is one to characterize the difference between emotions being strangled, throttled, smothered, stifled, choked or drowned? How much of the different imagery from the source domain is 'carried over' into these usages? Should the absence of rigid differences in metaphorical extensions like these not imply that the distinctions in the source domain may not be so neat either? I argue that all the observed differences do stand up to close examination, since they are reflected in the language itself, given the different collocational patterns that have been observed and on which the distinctions drawn up here have been based. For instance, while it may be true that loanwords enrich the expressive power of a language, it is equally true that after a while, when these loans have become fully incorporated into the system of the language, they start to lead a life of their own and develop new meanings. The analysis of authentic data has shown that although smother and stifle were certainly largely equivalent in the early Middle English period, they have crystallized into different prototypical centres as reflected in prototypically different types of causes, i.e. coverage and atmospheric pollution. Further, smother and drown show some overlapping, but this occurs in usages which are atypical for both verbs and it appears to be motivated by differences in the conceptualization of the degree of coverage. At the same time, it cannot be denied that even in contemporary English there is still some degree of overlapping. As explained in the introduction, this may be due to the ontological proximity. However, an ontologically identical base can be conceptualized in different ways, and especially the variable conceptualization of the actual cause (e.g. smoke as gaseous substance or as an airtight cover) may lead to some semantic overlap within the field of SUFFOCATE verbs. In short, it can be reiterated that the differences presented above show the prototypes for these verbs, i.e. cognitively salient polarizations of the

176


differences, which, however, may be blurred at the periphery and be neutralized in mutually overlapping extensions. Turning then to metaphorical extensions, how much of these differences in the source domain can be said to be carried over in metaphorical mappings? Clearly, also for these usages, no hard and fast distinctions can be drawn. Nevertheless, collocational patterns and contextual features indicate that some of the distinctions of the source domain are preserved in these usages as well. Let us, by way of example, reconsider the extension to the suppression of sound, already illustrated above. Our data suggest that choke is typically used when the suppression of sound is self-imposed or brought about by emotions, which lines up with the internal perspective typical of this verb, enforced by the typical occurrence of with expressing emotional saturation of the Affected. It can be noted that this metaphorical extension to the domain of speech acts is quite common for choke, so much so that the verb is often used as a full-fledged speech verb, e.g. "I've always wanted to meet you, " Jason choked (CASS). Smother and drown out typically occur in contexts where sounds are suppressed either through physical coverage of the sound source or by other, louder sounds, in which the metaphorical conception of one sound enveloping the other is activated (see the next section for an analysis along these lines for drown out). The metaphorical extension is also quite frequent with stifle and in this usage, it is quite close to choke and smother. However, next to drown, it is the only verb which has been attested with sounds that do not originate from the throat, as in a stifled thump. The explanation for this may be that both stifle and drown attribute the inaudibility of the sound to the muffling environment (compare The fog...stifled the roar of the London traffic). The focus on the environment is a feature retained from the literal use οf stifle. The low salience of goal-achievement that was said to characterize throttle has percolated to the metaphorical use as well: the phrase a smooth, somewhat throttled basso refers to a bass continuing to sing but using an imperfect technique resulting in a tense throat. In contrast, strangle seems to impose a more abrupt stoppage of sound, often due to an external constriction of the throat, e.g. He clutches his throat (with a strangled cry) "My God!" (LDC). In short, while not absolute, there is clearly a tendency for some of the prototypical features of the verbs to be retained in their metaphorical uses. The analysis showing different collocational patterns for the SUFFOCATE verbs can be complemented with an investigation into constructional variation. In tandem with the lexical differentiation, the constructional possibilities of the SUFFOCATE verbs have further crystallized to more clear-cut distinctions. For instance, the internal perspective implied by verbs such as choke· or drown or the focus on the sensation prominent with suffocate and asphyxiate is reflected in


177

the prototypicality of an ergative construction. In some uses, a non-effective is the most common (e.g. he drowned in the lake) or even the sole possibility (e.g. the kitten choked on the food). These usages strongly focus on the activity on behalf of the Affected, and hence their typical association with the non-effective construction. Such construals are diametrically opposed to the transitive ones with strangle, throttle or smother which impose onto the event an AG-PROC-AF schema which centres around the Actor and demotes the activity of the Affected even though in the world 'out there' it is known to co-participate. Yet it was also shown that a typically ergative verb such as choke extrapolates away from the blockage and shifts the focus to external constriction, an effective construction appears to be most typical. To conclude, then, it can be said that the high degree of lexical and constructional convergence occurring in the 16th to early 19th century has been reverted to what appears to be their etymological prototypes. Even if not always clear-cut, the differentiation in contemporary English is borne out by constructional and collocational patterns and contextual clues. In any case, the categories should not be thought of as rigidly delineated, but as prototypestructured categories with blurred edges, i.e. grey transition areas between the different verbs. The above discussion has clarified that these areas may be stronger for some verbs than for other, leading to close associations. This semantic proximity leads to different clusters within the group of SUFFOCATE verbs. The subgroups suggested by the data can be represented as in the following simplified overview, in which the common feature is specified in SMALL CAPS. The brackets indicate that an item is less prototypical, either in general or for the transition area in particular. (i)

strangle and {throttle): EXTERNAL CONSTRICTION

(ii) (iii) (iv)

drown and smother: (PHYSICAL) COVERAGE (smother) and stifle: (ATMOSPHERIC) COVERAGE suffocate and (asphyxiate): SUFFOCATION

The verb choke stands somewhat apart within the group given its salient reference to swallowing the wrong way and its semantic flexibility, even in contemporary English, which makes that it often encroaches on the semantic space of all the other SUFFOCATE verbs. While the main lines of the present description have been distilled from linguistic data itself, and are thus to a large extent, true to the facts, additional research is warranted, both corpus analysis and psycho-linguistic investigation, to further substantiate the cognitive reality of the lexical and constructional convergence and/or differentiation within the group of SUFFOCATE verbs. Such an analysis would also paint a more accurate picture of dialectal (or even idiolectal) variation which, as I have come to

178


understand from comments by individual speakers, appears to be relevant to these verbs. 6.4 Transitivity andforce-dynamics: phrasal verbs 6.4.1 Overview of constructional deviations One of the frequent constructional subtypes which has not yet been discussed is the verb-particle combination, as in choke {down/back/out/off}, and drown out. The reason why they have not yet been dealt with is that they all share an important feature: they no longer allow the ergative non-effective construction, as illustrated by the following examples. (88)a. when he had choked down the last detestable bite (CASS) b. *the last detestable bite choked down. (89)a. Choking back sobs, he remarked, "It's only a game, right?" (WSJ) b. *Sobs choked back. (90)a. I choked out some words through my grief and misery (NOV) b.*some words choked out. (91)a. that rate may be high enough to choke off economic growth ... (WSJ) b.*economic growth choked off. (92)a. The activists ... sounded a horn to drown out the opening bell (WSJ) b. *The opening bell drowned out. While the unacceptability of the b-sentences may be intuitively clear, there is still the problem of explaining the phenomenon within the present framework. The impossibility of the ergative non-effective clearly illustrates that the ergative model is no longer activated. What is it that takes these prototypically ergative verbs into the transitive paradigm? I argue that the transitive character of the verb-particle construction is in each case due to the merger of the semantics of verb and particle, in which the increased force-dynamics of the conceptualization plays an important role. In many ways, our views on the semantic purport of particles align with those expressed in Rudzka-Ostyn (1988), although she does not consider how particles influence a verb's constructional possibilities. To clarify this point, let us briefly consider some possibilities attested with the SUFFOCATE verbs. 6.4.2 Choke down Choke down is a metaphorical extension mediated by metonymy: it is characterized against the background of the domain of eating, and no longer that of suffocation. The extension entails a metonymical shift from the windpipe's orifice to that of the oesophagus, which is unproblematic given their physical proximity in the throat as well as the common phenomenon of


179

swallowing the wrong way while eating. Further, both the conceptualization of swallowing the wrong way and that coded by choke down saliently incorporate the notion of retching, since for the latter there is the strong implication that considerable physical effort is involved. The effort can be attributed to various reasons, such as the food's shape, its quantity, or (most commonly) its inferior quality or revolting taste, as explicitly coded by detestable bite in example (88). The usage of choke down in reference to swallowing with great difficulty may in turn provide the source for additional metaphorical extensions, as in (93)

Jim choked down the director's rudeness because otherwise he might have lost his job (LDPV).

The construction choke down drastically changes the conception of causality of the source domain. It specifies that great difficulty or force was involved in the act of swallowing since the direction of swallowing goes against that of the bodily reaction of expelling the 'uneatable' object (which need not be inedible). In terms of Talmy's force-dynamics (1985a), we can say that the eating entity energetically interacts on the object to be swallowed and overcomes its tendency to upward movement (itself triggered by the retching). The enhanced forcedynamic interaction converts the typical ergative constellation of the base verb into a transitive construal with a shift of participants. The Medium of the base process, the choking entity, becomes the Actor that energetically impinges on the uneatable object, which as a result is forced to move along the downward path of the oesophagus. The Actor may retain from the prototype of choke some degree of affectedness; swallowing down something that one's body tries to expel can in many ways be a painful experience. 6.4.3 Choke back The phrasal verb choke back is fairly infrequent and mostly occurs in the meaning where sobs or other expressions of emotion are being suppressed, as in example (89) above. The differences with choke down pertain to (1) the domain relative to which the expression is characterized, (2) the extension of the conceived path, and (3) the entity traversing this path. Choke back refers to the halting of the emission of breath, possibly accompanied by vocal sound. In the conception designated by choke back, stopping the emission of breath and/or vocal sound equals reversing its conceived forward direction (which aligns with the experientially grounded conception of the body as having an inherent frontback orientation).58 The usage is metaphorical, since what is retained is the effort needed to control one's emotions, or at least, the expression of them (most conventionalized is choke back sobs). Like with the particle down, the See, for instance, Lakoff & Johnson (1980) and Johnson (1987).

180


resulting construal affects the participant constellation and has become fully transitive. The Medium of the ergative prototype becomes a transitive Actor who exerts his influence and control on a Goal, in this particular case prototypically the expression of one's emotions. There is an interesting attestation in the LDC, in which the notion of emission extends to another physical domain, namely that of delivering a baby: (94) BRI: SHEILA:

BRI:

Sheila's got a theory about Joe's birth. She doesn't blame the doctors. She blames herself. I don't say that. I say it wasn't entirely the doctors. It was because she choked it back. (LDC)

Clearly, this usage only retains the force-dynamic physical reaction targeted at reversing the normal expulsion from the body. 6.4.4 Choke out Also characterized against the background of the conduit metaphor, choke out denotes exactly the opposite of choke back: considerable physical effort is required of a speaker to successfully speak (i.e. get the words out). As with choke back, it is typically the case that the speaker's emotional state prevents his normal capacities for speech, as in example (90) above (7 choked out some words through my grief and misery), yet this need not always be so: (95)

The Chinooks ... seemed to choke out their words (OED)

This usage implies a condescending view on primitive languages as having an inarticulate sound. However, the following usage shows that choke out can involve a more salient reference to the source domain: (96)

he choked out the salty taste which swept through his windpipe. (CASS)

Typical of a coding with choke, there is a strong contextual profile on the internal organs affected in the process {windpipe). Choke out retains from the source domain the salient evocation of the strong physical reaction of trying to expel the unbreathable substance (ignoring the metonymical use of taste in the above example). It is perhaps more this strong bodily reaction rather than the semantic import of the particle which is responsible for the force-dynamic character of the construction. The particle, in fact, externalizes the usual path, both for the unbreathable substance as its metaphorical counterparts (e.g. words flow out effortlessly). Somewhat surprisingly, the force of expulsion need not always come from within, as witnessed by the constructions like to choke out someone's life or to choke the life out of someone, in which the Instigator is


181

usually a human being constricting the throat from without. Notice that these usages (but also to choke the life from someone) involve a metonymical extension from the prototype, shifting the profile from the living entity to life itself (cf. Chapter 3, section 3.1). 6.4.5 Choke off Whereas with choke down, back, or out the entity into or out of which something is forced is still reasonably straightforward (by and large, the body), this is less so for choke off Moreover, how is the prototypical locative meaning of off(e.g. come off the roof at issue in this usage, given that in combination with choke it comes to emphasize total destruction or disappearance, as in example (91) above (to choke off economic growth)? Although this is not the place to elaborate on the semantics of the particle off its emphatic nature can be related to its more basic locative meaning "away from a certain location" (compare off-shore, off you go, lift the cup off the table, etc.). Expanding Lindner's (1982) analysis, I argue that in the prototypical conception designated by off one entity, the trajector, is initially situated on another, the landmark, specified as a bounded region of some sort. The trajector follows a path such that it changes its position vis-à-vis the landmark, until eventually, it is no longer supported by the landmark. Following Lindner's (Langackerian) diagrams, the locative meaning of off can be diagrammed as follows:

Fig. 20: The extended locative relation designated by off This figure portrays a series of locative relations, the leftmost configuration is initial and the rightmost is final. Following Lindner this meaning of off can be characterized as an extended locative relation: a relation because it profiles two entities relative to each other, locative because it is defined in the spatial domain, and extended because it is a series of configurations representing a path. Now, any state, including existence, can be conceptualized as a concrete object or a location: to exist is to exist in locations.59 Consequently, off may extend metaphorically to designate a relation in which the location represents 59

Cf. Lakoff & Johnson 1980:29ff, Lakoff 1987:518, Johnson 1987, Vandeloise 1991.

182


the abstract region of existence. When combining with verbs of destruction which by themselves mean that some existence is halted, off emphasizes the total annihilation, as elucidated by the following examples: (97) (98)

the primary reason for Auschwitz was to attempt to wipe the Jewish race off this earth. (WSJ) ... some of the nation's poorest black citizens believe that the ghetto crack wars are part of a white plot to kill them off. (WSJ)

Not incidentally, the event in the second example is also expressed in terms of genocide and eliminate. Note, however, that there is a clear difference between the two examples. The first example still profiles the location {earth) from which the Jews are removed (this region is the landmark of the relation designated by off). In the second sentence, this region remains fully off-stage, which also downgrades the image of abstract motion, whereas this is still undeniably present in the first example. The reason of course is that wipe is prototypically a verb of motion, whereas kill, like choke, is not.60 As with the other particles, the combination with off effects a higher prominence of the force-dynamic component: converting the on-configuration into a off-configuration requires considerable energy, and this energy may either originate within the entity itself, i.e. it has "an intrinsic tendency toward manifesting [force]" (Talmy 1985a:297), or the energy comes from a second entity that overcomes the supported entity's "intrinsic tendency toward rest" (1985a:297). It is the latter configuration that holds for choke off Once more, the prominence of the force-dynamic component affects the conception of causation: a non-effective construction with choke off is presumably unacceptable: *Economic growth choked off I did, however, come across two noneffective constructions: (99) (100)

Tom choked off and whispered: "Look, Hucky, look! ..." (NOV) The words choked off in his throat. (OED on CD-ROM, 1970)

On the total of attestations these two seem like isolated cases (note that both pertain to speech events), yet they indicate that the ergative model may still be activated. Undeniably, however, the data show that choke off strongly tends to instantiate an effective construal. Mostly also, the Instigator is either a less welldefined situation or is left unspecified, which is quite common with ergative construals. A final element concerns the syntagmatic combination that characterizes choke off: the contemporary data present conceptualizations where the 60

See Talmy (1985b) and Levin & Rappaport (1991, 1992) on verbs of motion and their constructional possibilities.


183

affected entity is a process (e.g. economic growth or nationwide recessions). This may suggest that choke off is gradually specializing towards a meaning paraphrasable as "to halt the progress of, to stop". It is perhaps from this viewpoint that the non-effective construction in (99) and (100) should be related to a re-interpretation which has led to profiling the moving entity (the words) instead of the process. However, they remain relatively exceptional uses, and it is to be doubted whether this reinterpretation will extend to other uses as well. 6.4.6 Choke up In contemporary English, a more unmarked alternative for the non-effective choke off construction is the phrasal verb choke up, which often evokes the concept of being paralysed because of strong emotions. The addition of the particle up does not have the same effect on the effective construal as do the other particles, and an ergative conception of the process designated by choke up is still possible, as illustrated by the following example: (101)

"The incu — the —" Leach choked up a bit, but finally managed to say "The incubator blew up on them" (CASS)

While the past participle choked up is the most common construction, primarily in reference to someone in an extreme emotional state, example (101) shows that it is still possible to cancel out the force-dynamic component in a non-effective construction. This type of construction is well-motivated in view of the abstract path that is entailed by the particle up. In contradistinction with the particles down, back, and off the congestion is conceived as lying ahead of the assumed vantage point. What motivates such an internal vantage point with choke up? In their analysis of the concept ANGER, Lakoff & Kövecses (1987) argue that our conceptualization of emotions is in large measure structured by the general metaphor "the body is a container for the emotions".61 In this conception, emotions can be viewed as substances filling that container, which may lead to 'combustion' (e.g. he's filled with anger, she's about to explode, etc.). The particle up in choke up is motivated by such a conception of upward motion. What then is the contribution of choke? The image retained from the source domain is that the throat, from which vocalizations originate, is blocked. As such, the construction is similar to be choked with. While typically situated in the domain of speech acts, extensions to non-verbal domains are possible (e.g. the traffic is choked up).

61

See, however, Geeraerts & Grondelaers (1991) for a cogent critique of Lakoff & Kövecses's analysis.

184


6.4.7 Drown out A last illustration of the transitivizing effect of a particle is drown out, which prototypically refers to the suppression of sound, usually by another, overwhelming sound. The imperceptibility of one sound because of another is based on our everyday auditory experience and also the conceptualization in terms of drown is well-motivated. Underlying this usage is the conceptualization of sound as a liquid substance (cf. also sound waves or a torrent of words), which can engulf or overwhelm an entity. (Note that overwhelm itself is an extension from the original meaning of the (now archaic) verb whelm "being covered by a hollow vessel, being covered or submerged in water".) Sounds are commonly thought of as mingling like water to form a homogeneous, perceptibly indistinguishable mass. If sounds are being suppressed by entities of a different type (i.e. not by another sound), drown is less likely to occur and other codings will be more common, like muffle or smother which evoke a conception of padding the sound emitting object. The total disappearance from auditory perception is what motivates the use of the particle out, whose meaning can again be satisfactorily explained drawing on Lindner (1982). In the prototypical conception designated by out, one entity, the trajector, is initially situated within the confines of the other, the landmark, specified as a bounded region of some sort. The trajector follows a path such that it changes its position vis-à-vis the landmark, until eventually, it is no longer situated within the landmark. Like off, out designates an extended locative relation between container and contained. In a construction like drown out, the particle out specifies that the embedded trajector (the 'drowning' entity) is no longer within the confines of the conceptualizer's "range of cognitive and perceptual access" (Lindner 1982:311), the abstract region that functions as the implied landmark. In other words, the trajector is conceptualized as having become inaccessible to perception, it has moved 'out of existence' (cf. also Talmy 1991). The conceptualization of one sound drowning out another can be represented as follows:

Fig. 21: Diagrammatic representation of X drowns out x


185

The situation involves the integration of two relations: the first relation is that designated by drown, and holds between a larger sound (S; tr1 that acts upon a smaller sound (s; lm1). The second relation is that designated by out, picturing the smaller sound (s; tr2) as moving out of the domain of auditory perception (lm2). The integration of course pertains to the causal relationship: it is the action of S that makes s disappear from auditory perception. Although with drown out the auditory domain has become conventionalized, the pattern is not restricted to it. Consider the following examples: (102) (103)

Deserted cottages, whose tenants ... had been 'drowned out' (OED, 1894). ... the bright light ... drowning out everything in its dazzling brilliance (CASS)

In the first example (a now uncommon usage), the hedges indicate the metaphorical usage. The landmark of out is the location from where the inhabitants are driven, i.e. deserted cottages. The OED also lists uses where inhabitants are driven from their habitat by actual flooding, in which the literal and the metaphorical are intertwined. In the second example, the domain is that of visual perception. Like sound, light is also commonly conceptualized in terms of water or waves. Interestingly, the earliest OED-attestations for the metaphorical conception of a sound and light being drowned by another sound or light indicate, by their explicit character, the novelty of the metaphor in the early 17th century: (104) (105)

The sound ... was drowned by some louder noise. (OED, 1605) A great Light drowneth a smaller, that it cannot be seene; As the Sunne that of a Gloworme (OED on CD-ROM, 1626).

Such effective construction without the particle out are still possible in contemporary English. Moreover, as for the literal usages, agentless passives are also common in this usage. In these constructions, the overpowering entity may surface in an in-adjunct, expressing the idea of (abstract) containment. This may lead to some nice contrastive pairs, as in: (106) (107)

Neville's cries were drowned out by the applause. (CASS) ... the ambassador's last words were drowned in applause. (CASS)

Although it is perhaps in principle possible to recast the sentences in a noneffective construction, the cries/words drowned in the applause, the result is a marked construction, as also signalled by its absence from all the corpora consulted (including the OED). Thus, when drown alone is used in the meaning

186


of "remove from the conceptualizer's range of access by overpowering" an effective construction with an overt Instigator is most prominent, most likely due to the prominence of this more powerful Instigator in our experience. While a non-effective construction might still be feasible in these cases, it is ruled out altogether with the phrasal verb drown out. The addition of the particle pushes the construal potential clear into the transitive realm. It be noted that in contemporary English, the addition of the particle out is most common in this meaning; it occurs in 80% of the cases. Finally, it can be noted that the common metonymical shift from sound to the object emitting the sound may lead to a different participant constellation, in which the 'emitters' are profiled as directly participating in the process, as in (108)

Victor Laszlo leads Free French sympathizers in "La Marseillaise" to drown out the Nazis (WSJ).

Profiling the 'emitter' does introduce some amount of ambiguity. For instance, the larger context of (108) lacks any specification as to whether the Nazis are emitting any noise (singing, scanting slogans, etc.). The sentence may be interpreted as indicating that everything the Nazis stand for is ousted. 6.4.8 Summary In conclusion, the addition of particles which enhance the force-dynamic component of the construal appears to shift the typically ergative verbs choke and drown to the transitive paradigm. With the particles down, back, off and out, there is the clear implication that considerable effort is required of the Agent to control the (abstract) movement of the Affected. The impossibility of the non-effective indicates that for these constructions the ergative model is no longer activated, but that they realize an AC-PROC-GO construal. In some cases, such as choke {back/down}, there is also a change of participants since the Medium of the ergative base process does not become the Goal (as would be expected in the case of a normal transitivization) but the Actor who targets his action at some other object, the cause of the (possibly metaphorical) retching. The present analysis confirms that particles may affect the transitivity of the clause, which has been observed before. Hopper and Thompson (1980:261-2) and Hopper (1985) point out that particles in English may increase the transitivity, often amounting to a difference between fiilly affected patient (verb with particle) and partially affected patient (verb without particle).62 However, as Hopper (1985) correctly notes, "the difference [...] is not a consistent one, and sensitive to individual verbs and particles". This is confirmed by the findings 62

Cf. also Talmy (1985b) for pertinent observations concerning particles, which he situates in a larger class of "satellites" (which also includes affixes, for instance).


187

regarding choke up and its continuing reliance on the ergative model. No doubt, it is not just the particles in isolation that change the conception of causality, but the conventionalized, semantically motivated integration of verb and particle. The present analysis also nicely lines up with observations concerning outprefixation that have been formulated in the literature. For example, Keyser & Roeper (1984), drawing on Bresnan (1981), point out that the out-prefixation converts a one-participant construction (an intransitive, an ergative noneffective, or an objectless transitive) into a transitive effective one: (109)a. b. c. (110)a. b. c. (111)a. b. c.

John ran. (Keyser & Roeper 1984) John outran Bill. *John outran. The basketball bounced. (Keyser & Roeper 1984) The basketball outbounced the baseball. *The basketball outbounced. John killed. (Keyser & Roeper 1984) John outkilled Fred. *John outkilled.

Significantly, with outbounce the ergative non-effective is no longer possible. Conversely, all can occur in a transitive middle construction, e.g. BillGo {outruns/outkills} easily or BaseballsGo outbounce easily. Being fully transitive, the latter is not ambiguous in the same sense that this window opened easily is. When particles combine with transitive verbs, e.g. kill off, the increased transitivity is not as marked as with prototypically ergative verbs. Yet, in addition to the lexical reinforcement of the destruction, the impact of the particle may be reflected in the marked character of the objectless transitive (similar to the situation for out-prefixation): (112)a. Soldiers trained to kill. b. ??Soldiers trained to kill off. The reason for this is that kill off saliently evokes the concept of a fully affected Goal; the limiting profile on the action component which characterizes the objectless transitive seems therefore to be excluded. 6.5 Conclusions This chapter has illustrated how the ergative predilection of the SUFFOCATE verbs is motivated (but not absolutely predicted) by our experience of the world. Some factors that have been argued to motivate an ergative event conception are the salient bodily activity of the Affected, the prototypical imperceptibility of the cause mostly seen as internally affecting the Medium, and the common temporal distance between cause and effect. In this view, the

188


ergativization affecting the transitive SUFFOCATE verbs in the (Early) Modern English period finds a natural explanation. Why the ergative pattern did not entrench for strangle, throttle and smother requires more comparative research than can be afforded here. Plausibly, the salience of an external cause incorporated in the conceptual structures of these verbs can be held responsible for the non-entrenchment of the ergative pattern. The internal/external distinction with respect to the cause appears to motivate not only the distinction between the transitive and the ergative SUFFOCATE verbs but also the relatively systematic occurrence of an effective construction with prototypically ergative verbs when the cause is overtly external. The lexical and constructional flexibility that characterizes the evolution of the SUFFOCATE verbs should also be seen against the background of the ontological proximity of the conceived events. Undoubtedly, a crucial factor encouraging the convergence is the variable conceptualization of one and the same cause. Such variation especially occurs in the case of less 'tangible' causes, like gaseous substances or smoke or the like. This gave rise to a high degree of interchangeability of the different SUFFOCATE verbs for a period of over 500 hundred years, during which the verbs, with varying success and degrees of entrenchment came to encroach upon each other's conceptual space. In the evolution to present-day English much of that convergence has been lost, and the verbs have crystallized around more differentiated prototypes, mostly a reorientation towards their etymologically basic meaning. Indisputably, there is still some degree of overlapping, e.g. between drown and smother, strangle and throttle, or smother and stifle, although even here careful data analysis has pointed at certain subtle differences. The differences discerned are obviously not rigid but represent prototypical tendencies as reflected in different syntagmatic combinations or different contextual implications. The semantic proximity of the SUFFOCATE verbs is also reflected in, and further enforced by, 'shared metaphorical extensions, especially the frequent metaphtonymical mapping to the domain of speech acts. In general, the enormous flexibility that has characterized the evolution of the SUFFOCATE verbs supports our claim that lexical and constructional meaning should not be seen as strictly separate but as dynamically interacting aspects of clause construal. More particularly, our description has shown that for some usages of a verb, a particular type of construction is prototypical or even obligatory. To recapitulate a few examples, recall the systematic use of the noneffective construction with choke when used in reference to swallowing the wrong way, the use of an ergative effective construction with choke when the cause is clearly external, or the prototypicality of the non-effective with drown in reference to a drowning event. The ergativization of throttle down an engine


189

markedly stands apart in the item's lexical and constructional profile, so much so that it seems to tilt its prototype to a completely different domain altogether. The phrasal verbs clearly show how the constructional potential of a verb is determined by its usage. In constructions with particles that enhance the forcedynamic component in the process conception, the transitivity of the construction is augmented and, in the case of prototypically ergative verbs like choke or drown, the non-effective construction is no longer possible. In other words, because the particle saliently refers to the force-dynamic component, this component can no longer be cancelled out. The semantics of the ergative noneffective construction, viz. that the process occurred spontaneously, without an agent, conflicts with the semantics of the verb-particle combination. The very structure of this chapter on the SUFFOCATE verbs reveals that these verbs have an entirely different character than the MURDER verbs, which were shown to be much more stable in their usage and their Agent-centredness. The SUFFOCATE verbs in contrast display an enormously rich and dynamic history of lexical and constructional extensions. For obvious reasons, the description in this chapter has been limited to some major patterns and evolutions. Nevertheless, I believe that it has revealed some of the complexities of the interplay between a verb's meaning and the construction in which it can occur. The next chapter will further substantiate this point by taking up the verb abort which is a more pronounced blend of transitive and ergative characteristics and whose behaviour is motivated perhaps even more systematically than that of the SUFFOCATE verbs. Moreover, the factors responsible for the blended constructional profile will turn out to be quite different.

CHAPTER 7 COMPLEXITIES OF THE TRANSITIVE-ERGATIVE INTERPLAY

7.1 The experiential basis of the transitive/ergative interplay As far as its conceptual content is concerned, abort is a marginal member of the field of killing.63 Whether abortion is to be considered an act of killing has been the topic of many a heated debate, which further influences the verb's already uncertain categorial status. In lexicological terms, the uncertainty involves the hyperonymy of the KILL category vis-à-vis ABORT. As can be expected in cases of clear hyperonymy, an abort-event may be named by kill (as well as murder, slaughter, etc.) but not vice versa. The vigorous debate on abortion is about the question whether the fetus is a living person or an entity that has human potential. Medical progress and the ensuing ethical debate about abortion have had an indubitable effect on the verb's linguistic structure, which has gradually shifted from designating a spontaneous abortion to designating a non-spontaneous one. Moreover, the ideological position of the speaker influences the choice of participant profiled as the affected entity. Equally determined by the speaker's ideological stance is the lexical variation in reference to an abortion event. The conception of causation has also been influenced: in contemporary English, particularly as represented by the WSJ corpus, abort has mixed transitive and ergative characteristics. What is more, the characteristics appear to parallel the distinction between the literal and metaphorical uses: the former are transitive, the latter primarily ergative. In short, also the ambivalent nature of abort appears to have a clear experiential basis, as it is motivated by the way in which humans have come to interact with the world, especially when it comes to medical care. More explicitly than with other verbs of killing, the experiential basis certainly has a rich ideological component, although this was shown to be important for some of the MURDER verbs as well. Since the ambivalent nature of abort is much more pronounced than with some of the SUFFOCATE verbs (e.g. smother), a more detailed investigation, 63

The present chapter revises and thus supersedes earlier analyses presented in Lemmens (1994, 1997).

192


especially into the diachronic evolution of the item, is called for. Such an investigation will once more present a concrete illustration of the ergativization process that has characterized Modern English. At the same time, the discussion provides evidence in support of my central hypothesis that the meaning of a verb influences its constructional possibilities. To have a good understanding of the situation to be accounted for, let us first have a look at the contemporary situation for abort. 7.2Lexicaland constructional variation in Modern English The description of the Modern English abort-cluster is almost exclusively based on the WSJ corpus, as the other text corpora yield only a handful of attestations, none of which were substantial enough to make a valuable contribution to the discussion. The formal distribution of the abort-cluster in the WSJ corpus is presented in Table 9.64 LEXICAL ITEM

FREQ

PCT

abortion aborted abort anti-abortionist abortive abortifacient aborting pro-abortionist

641 19 17 14 7 1

91.4% 2.7% 2.4% 2.0% 1.0% 0.1% 0.1% 0.1% 100.0%

TOTAL

1

1 701

Table 9: Frequency of abort in the WSJ corpus The high number of attestations (over 90%) of the noun abortion is no surprise, given the controversial status of the issue of abortion, a controversy that has repercussions at various levels in the United States (politics, jurisdiction, scientific research (e.g. fetal tissue research), health care, etc.). The Wall Street Journal, like any newspaper, devotes numerous articles to this issue, mostly using the nominal form. We should be careful with the interpretation of the high occurrence of abortion, yet the use of the noun does have its impact on the further development of the cluster. Regarding usage, it should be pointed out that all occurrences of abortion and the majority of the verb forms refer to a "non-spontaneous premature termination of a pregnancy". Similarly, the single attestation of abortifacient refers to the drug RU-486, which induces a premature termination of a pregnancy. Pro- and anti-abortionists are people in favour of or opposed to The description will show that it is relevant to distinguish the past participle from the verb.

COMPLEXITIES OF THE TRANSITIVE-ERGATIVE INTERPLAY

193

abortion. All these different forms instantiate, with their own particular profile, the prototypical meaning of the cluster, namely "to cause a premature expulsion of the fetus from the womb, thereby preventing the pregnancy to successfully culminate in the birth of a mature baby" {abortus provocatus or sometimes also called criminal abortion). The following examples from the corpus represent some instantiations of the prototype, illustrating at the same time the possible metonymical variation. Since the type of Affected entity is crucial in these examples, it has been highlighted in italics. (1) (2) (3)

Prenatal testing ... might motivate some pregnant women to abort fetuses afflicted with cystic fibrosis. (WSJ) ... black women are more than twice as likely to abort their children as white women. (WSJ) ... some 400,000 black pregnancies are aborted each year. (WSJ)

The metonymical variations in profile that may occur in the literal domain are similar to those that have been discussed more elaborately in Chapter 3, section 3.1 More specifically, the relation between (1) and (2) is a Part/Part metonymy: the profile shifts from the developing to developed entity. Example (3) extends to profiling the process itself, which is a clear instance of a Part/Whole relationship. This is where abort diverges from kill: in literal usage of kill a profile on life as a process is impossible (e.g. *The burglars killed my dog's life). As example (3) shows this is unproblematic with abort. The explanation I take, is that the process, a pregnancy, has fixed boundaries and one knows when in the normal course of events the pregnancy will successfully terminate. Collins Cobuild (1987) even lists the doctor aborted the woman as a possibility, but some people may find it odd or will reject it altogether. It is not attested in the WSJ corpus, but I did find a similar construction in postings to (unmoderated) newsgroups on the Internet (e.g. Helen and other aborted women failed to get an informed consent law enacted). It could be regarded as equivalent to a construction like kill the estuary where the location is in profile. At the end of this chapter, we will be in a position to explain why the metonymical shift to profiling the woman has become possible.65 The metonymical variations can be summarized by Figure 22, an adaptation of the diagram presented for kill (see Figure 11 above). Recall that the different states of the abort-process have been collapsed to one single box; the time arrow underneath the relational profile signals its temporal nature. The heavy lines within a box represent the profiled landmark. The item's most prototypical 65

I have noted that in Dutch, too, the phrase vrouwen werden geaborteerd "women were aborted" has come to be used more often. As with the English equivalent, many speakers object to it.

194


configuration (at least in the WSJ corpus, i.e. 75%), namely to abort a child, has been marked by the heavy-line rectangle. The construction to abort a fetus is less salient (12%).

Fig. 22: The metonymical extensions of literal abort Metaphorical extensions are quite common with the verb abort; for the past participle, the adjective abortive and the gerund, it is the only usage that has been attested. As can be expected, metonymical variations like those in the literal domain can also occur in metaphorical usages, as shown by the following examples: (4) (5)

...the company had aborted a probe into fraud and waste... (WSJ) The requirement "could have the unfortunate result of...aborting other technologies now under development," she said. (WSJ)

In (4), the thing affected by the action is an ongoing development, a probe, the outcome of which would have been, say, a report on its results; as such, it is the metaphoric counterpart of (3), in which the development is a pregnancy.66 In (5), the entity affected is not the development itself, but the thing undergoing the development, viz. the technologies (for safer air traffic) which, similar to developing fetuses in the literal counterparts such as in example (1), should have been able to grow to full maturity, to be used in airlines. In other words, examples (4) and (5) reveal that the metaphoric extensions also allow for a metonymical variation similar to that in the literal usage.

We will be using the term development rather than the more general label process to indicate that we are dealing here with a type of process in which an entity develops (changes) over time. In cases where the distinction between the two is no longer clear, the term process may also have been used.


195

In the WSJ corpus, the past participles always function as premodifier to a noun and occur only in metaphorical readings. Here are some typical uses: (6) (7) (8)

Aborted programs include decontrol of prices, sales of stateowned enterprises and strengthening the private sector — all of which seemed likely to move forward a year ago. (WSJ) The company...lost about $80 million ... on an aborted attempt to start a direct-broadcast satellite network. (WSJ) Mr. Guber...and Mr. Peters made an aborted bid to buy part of MGM/UA Entertainment Co. to run the MGM studio. (WSJ)

The past participle constructions are ambiguous, as they may mean "process that are aborted" or "processes that abort". The contextual presence of an Agent varies. For instance, in the report on the Chinese reform programs in (6) the context specifies that these programs were halted by the violent suppression of the students' protest march for more democracy (Beijing, Tianamen Square, June 4, 1989). For examples (7) and (8) there is only minimal implication that an Agent was involved at all, these uses seem to restrict the profile to the developing entity's failure to reach completion. The coding emphasizes that something went wrong in the course of the development, as a result of which success could not be achieved (in more biological terms, appropriate in the context of this chapter, the development could be labelled 'fruitless'). Nevertheless, some ambiguity is retained. The adjective abortive goes much further than the past participle, retaining only the 'unsuccessful' component and disregarding all other elements of the prototypical meaning of abort. (9)

After two abortive attempts at securing a foothold above the crucial 2.00 Deutsche mark level, the dollar sagged to finish softer. (WSJ)

Involved in this usage of abortive is a "metaphtonymy" (Goossens 1990). First, it maps metaphorically onto various kinds of developments or processes; secondly, within the metaphoric uses, the Part/Whole metonymy applies which profiles the development but not the entity under development. Typically, abortive but also aborted modify nouns that express a development or an event which often have a certain risk of going wrong, such as attempt, bid, invasion, rebellion. The alternation between effective and non-effective construals identifies abort as an ergative verb. There are, however, two problems. First, the data show that the non-effective is primarily restricted to metaphorical usage: a construction like the fetus aborted is marked and predominantly restricted to

196


medical jargon. Second, there is a problematic objectless transitive in literal use, as evidenced by examples(10) and (11): (10) (11)

Thus, a woman who used RU-486 to have an abortion would have to make three trips ... and a third trip a week later to make sure she has completely aborted. (WSJ) I assert that a mother has rights, and a baby does not. That being the case, the mother can abort, if she so chooses. (ABORTION, pro-choicer)

Normally, ergative verbs do not tolerate such an objectless construction; kittens choke cannot be interpreted as coding a conceptualization in which the kittens are the Actors affecting a choke-process within an entity left unspecified. Why, then, is this possible with abort? There are two possible options to account for the situation. One account is that abort, a loanword from Latin (a pronounced nominative/accusative language) has been subjected to an ergativization process with the objectless construction in (10) as a relic from the earlier transitive period. This account is plausible, as it has been shown in the discussion of starve, hunger and the SUFFOCATE verbs that such an ergativization has indeed been taking place over the past five hundred years. The alternative hypothesis is to assume exactly the reverse, viz. a transitivization process, postulating that abort which was originally an ergative verb has developed towards a transitive construction, with the objectless construction in (10) as its logical exponent. The disadvantage of this hypothesis is that many of the examples cited so far have to be regarded as relics of the older ergative situation, despite the fact that for the past participle and the adjective abortive an ergative conception is most productive, if not the sole possibility. Against this hypothesis is also that it posits a development going against the tide of 'mainstream' ergativization. A careful examination of historical data reveals that, strangely enough, both hypotheses hold for abort, with the understanding (1) that the processes of ergativization and transitivization have operated in different time periods and (2) that the transitivization selectively applies to the literal uses only. So, the present day semantic structure of the verb abort and its formal associates67 has been shaped by the interplay of the ergative and the transitive paradigms, which has left a somewhat unbalanced category. The literal uses of any of the abortforms, i.e. those uses that refer to the premature arrest of a pregnancy, all 67

It should be noted in passing that none of the other still existing forms (e.g. abortion or abortifacient) are derived from the English verb, but are original Latin derivations. Hence, I call them associates rather than derivations.


197

instantiate a transitive reading, in which someone or something deliberately causes the premature expulsion of the fetus. Only the metaphoric readings may in addition to an effective reading also realize a non-effective construction in which the development that goes wrong (which thus misses its purpose) is profiled and the cause is backgrounded. To elucidate the complex interplay of the ergative and transitive paradigms, the next part will present a more elaborate description of the diachronic development of the cluster. The discussion will illustrate how the specific semantic structure of abort realized, and in turn influenced these schematic grammatical patterns. I would like to reiterate that this historical reconstruction will have a qualitative rather than quantitative orientation since it is not based on an extensive corpus of representative historical data that might reveal, for different time periods, which meanings (or types of construction) were prototypical and which were peripheral. Nevertheless, the description is more than just speculation, since, first of all, the information in the OED and the Middle English Dictionary has been sufficient to outline the main tracks along which the cluster developed and secondly, the data fit the ergative-transitive interplay hypothesis so nicely that it would be unfortunate not to consider them. 7.3 Ergativization of abort in Early Modern English The attestations in the OED present the uses of the abort-forms in the 14th, 15th and 16th century as a mixed bag. This diversity can, however, be accounted for by an evolution that took place in Latin, from which English, in two different time periods, drew its items. It is clearly beyond the scope of this discussion to examine exhaustively the precise details of the borrowing processes, yet the Latin variation merits some thought. According to the OED, abort originates from the past participle form abortus of the Latin intransitive verb ab+oriri (literally, "away" + "come into being") meaning "to miscarry" or in figurative use, "to go wrong; to disappear". Most etymological dictionaries (e.g. Shipley 1945) and other dictionaries (e.g. Webster's 1973) confirm this etymology. Partridge (1966), however, says that the Latin past participle abortus was the basis for the formation of a new Latin verb abortare (meaning "to miscarry") from which English abort is derived. This verb is also referred to by the OED, but only as an element of crosslinguistic comparison. In a sense all these etymologies are correct, yet they raise more questions than they answer. Let us, therefore, take a closer look at the Latin data.68 The Thesaurus Linguae Latinae I (col. 125-127; henceforth TLL) shows that the earliest form is the deponent verb aboriri, which (in opposition to oriri) I am grateful to Willy Clarysse and Willy Evenepoel for their help with the Latin data.

198


means "to disappear, to cease existing". In its most common use, i.e. referring to an unsuccessful pregnancy, it assumes the fetus as the figure, i.e. the participant that is born too early in the pregnancy and as a result ceases to exist or fails to come into existence. The following sentences may serve as examples (the figure is indicated in italics): (12) ut fetus ... in corpore suo concepii aboriantur. (TLL) "such that the fetuses ... conceived in her body would be born prematurely" (13) ... qui imperfecti adhuc et de ventre aborsi sunt (TLL) "... that still are not yet developed and are born prematurely from the womb". Of course, the woman is also part of the conceptual cluster of aboriri, as the ground, and is often expressed periphrastically in the clause (cf. in corpore suo "in her body" and de ventre "from the womb" in the two examples above). It is precisely this proximity that triggered a change in the verb's usage into one in which the woman becomes the figure, the participant most affected by the event. Löschhorn (1976:111) sees constructions such as the following as creating the possibility of this evolution. (14)

... vinum si praegnans biberit, fieri ut Δ aboriatur. (TLL, Varro) "If (a woman who is) pregnant drinks wine, it might cause (the fetus) to be born prematurely"

In Varro's time (first century BC) the subject of aboriantur was undoubtedly the omitted fetus. A later syntactic reconstruction allows, according to Löschhorn, the implied antecedent of praegnans (mulier) to be taken as the subject of the verb. Löschhom's explanation is correct, yet his restricted view of the change as a pure syntactic phenomenon—which has the unwarranted extrapolation that most usages of aborori must have manifested a syntactic ambiguity of the type illustrated in (14)—misses the point. It is the conceptual rather than the syntactic proximity of the woman that can explain why speakers of late Latin came to a reinterpretation of utterances like (14). I argue that what is at stake here is an example of a metonymical figure/ground reversal: the woman, who was initially the ground (cf. de ventre and in corpore suo) becomes foregrounded as the figure in the relation designated by aborori. What triggered this reversal in Latin (and what later standardized the reversed situation) is food for thought. The noun abortio, a nominalization of aboriri (the -tio suffix refers to an action), may have helped in bringing about the reversal, as it does not explicitly profile any participant, e.g.

COMPLEXITIES OF THE TRANSITIVE-ERGATIVE INTERPLAY (15)

199

abortio quibusdam, non partus, videtur (TLL) "as it seemed to some people an abortion/miscarriage, not a (regular) delivery"

Nominalizations such as this one are characterized by some degree of ambiguity in the identification of the participants (cf. also many contemporary English nominalizations) and such an ambiguity may have had an impact on the evolution of the verb aboriri. In addition, socio-cultural factors may have played a major role in this evolution, but I will not probe into these here. What is important for the present discussion is that aboriri developed into a verb that, as unequivocally illustrated by the following examples, had the woman (or its metaphoric counterpart as in one of the examples) as the figure (it became the subject of the verb, indicated in italics). (16)

(17)

Babylonia cum alienis etiam suas aborsa est pereundo divitias (TLL) "With its destruction, Babylonia has, together with the wealth of others, also lost its own wealth." xvij matres [oves] fuerunt mortue ante partum et xxj stereles et xxiiij abortae (DMLBS)69 "17 mother sheep were dead before the delivery, 21 were sterile and 24 had a miscarriage"

This new use of the verb aboriri (to which Donatus (4th century) still depreciatorily reacts non est Latinum 'aborsa est', sed 'abortum fecit ' ) becomes conventionalized in the later forms abortare and denominative abortire (or deponent abortiri). The Medieval tendency to turn irregular verbs into regular ones (usually by changing them into verbs ending in -are) also serves, at least in this particular case, as a regularization of the semantic structure: the new verb forms can only occur with the woman as subject (cf. Hiltbrunner 1981:48ff).70 There are also attestations which take the fetus as the direct object, as in (18).

69

Aborsus and abortus are spelling variants of the past participle of aboriri. The attestations are drawn from Latham (1975) Dictionary of Medieval Latin from British Sources; the roman numbers are taken over as attested in the dictionary. 70 Mignot (1969:71-72) observes that abortire (or deponent abortin) semantically lines up with other verbs on -ire that express physical, mental, or moral defects (e.g. moriri "to die", dementire "to lose one's reason", ineptire "to be silly", etc.). Strikingly, it is the deponent abortiri which is most common in Latin documents written in England (next to some attestations of aboriri that have the woman as figure, cf. Latham 1975).

200


(18)

(19)

si percussint mulierem in utero habentem et abortavit non deformatum (TLL) "if they hit a pregnant woman and she prematurely delivered a child not yet formed"71 mulier que frequenter partitur in matrice de facile abortito (DMLBS)

(20)

"a woman who often delivers, will easily miscarry" sic mater ecclesia pariendo filios abortiri compellitur (DMLBS) "such that the mother, the church, while bringing forth offspring, is compelled to miscarry"

Although there is much more to say about the evolution of aboriri from a 'fetus-oriented' to a 'woman-oriented' verb, the foregoing discussion suffices to explain some of the peculiarities of the English attestations. It should be clear from the outset that it would be incorrect to assume that the Latin constructions were simply copied into English, since the way in which a situation is construed closely depends on what imagery is embodied in the linguistic convention of a language (cf. Langacker 1991b:294). However, some of the semantic features of the Latin source were transferred to English equivalents. One of the features taken over is the figure/ground alignment. The oldest abort-forms that appear in English are the nouns aborsum and abort(e), and abortive that can be a noun or an adjective. (21) (22)

(23)

... medicynez prouokyng aborsum i. dedechilde (OED, ?al425) " medicines provoking a miscarriage or a stillborn child" Or when I>e materie is makynge aborte i. castyng childe (OED, ?al425) "or when the substance causes a miscarriage or a loss of the child" ... Þe childir Þat es abortiues, Þaa Þat er not born oliues ... (OED, al300) "the children that are 'abortive', those that are not born alive"

These examples illustrate that the early forms take over the 'fetus-orientation' which characterized the etymologically basic meaning of Latin aboriri. Note that in the first two examples, aborsum and aborte are fetus-oriented, but the larger structures prouokyng aborsum and makynge aborte bring in an external cause as well, in both cases a medicine (compare with vinum "wine" in (14)). Such usages, obviously alluding to an abortus provocatus, already contain the onset of the later systematization into a full ergative paradigm. In the middle of the 16th century, the English abort-forms, and most notably the verbal forms 71

Note that Latin deformatum literally means "away from being fully formed" and is thus not fully equivalent to English deformed.


201

and their derivations {abort, aborted, and aborting) start to have the woman as figure, still continuing the meaning of a spontaneous abortion as its prototype. This change suggests that, some 250 years after the first occurrence of aborsum, there was a new import from the Latin source, where, however, the verb aboriri and its younger derivations abortare and abortire had meanwhile taken on the woman as figure. The hypothesis of a second borrowing gains in plausibility when we consider that in Latin documents written in England in the 14th century and later (cf. Latham 1975) the abort-verbs (aboriri and abortiri) have the woman as central participant; we may assume that there was a close 'interaction' between Medieval Latin and English. Consider the following attestations from the OED: (24)

Abhorsion is when a woman is delyvered of her chylde before her tyme. (OED, 1547) (25) Avorter, to abort, or when a woman goeth not hir full time. (OED, 1580) (26) What wilt thou give them? barren wombes (or aborting wombs). (OED, 1632) (27) This Spring the Queen ... aborted of a son. (OED, 1655)

Both the noun abortion (sometimes also spelled abhorsion, cf. below) and the verb abort make their first appearance in the 16th century. The introduction of the verb may have been influential in the further development of the cluster, since the older forms readily subsumed the new coding possibility. The introduction of the woman as a prominent participant allowed a harmonious resolution of both the older uses that have the fetus as figure and the newer ones that have the woman as figure. The resolution lies in an overall ergativization of the conception of the abort-event. At its most schematic level, this ergative conception involves two participants, the most central of which is the Medium, the entity affected but also co-participating in the event. The other participant is the second energy source; it is the Instigator of the event. Given that in the 16th century and later, the literal use of abort mostly referred to a spontaneous abortion, it is only natural to arrive at a conception of the event which transcends the control of the woman and finds its primary energy source in the fetus. However, it was not until the 19th century that the ergative paradigm became the standard coding. Nevertheless, although not yet fully attuned to each other, the 16th and 17th century abort-forms can all be considered as instantiations of an ergative system, each of course with their own particulars of construal. In the 17th century, the codings still tend to alternate between taking either the child or the woman as figure (or their metaphoric counterpart, as in (29)), yet they

202


sometimes also combine the two in varying degree of explicitness. By way of example, consider some possibilities for the adjective abortive. (28) (29) (30) (31)

These abortive brates that are thus hastely brought into the world (OED,1614) Let it make thee Crest-falne, I, and alay this thy abortive Pride. (OED, 1593) Witches ... make Women abortive not to conceive. (OED, 1621) We but saue Our Innocent bodies from th' abortiue grave. (OED, 1611)

Although used metaphorically, the last usage implicitly relates the two participants in an ergative construal: death (the Instigator, metonymically referred to by grave) will compel us, innocent bodies (the Medium), to die. The external instigation that was explicitly acknowledged in earlier constructions with provoke or make, has now become part and parcel of the ergative paradigm as instantiated by the abort-forms. Quite explicit in such an Is-PROCME construal is the (now obsolete) verb abortive: (32) (33)

This is that which abortives the Perfection of the most glorious and useful Undertakings. (OED, 1699) When peace came so near to the birth, how it abortived ... comes now to be remembered. (OED, 1692)72

In the 17th century, ergative metaphors such as these also occur with the verb abort, as in (34)

[the Parliament] is aborted before it was born. (OED, 1614)

In its literal use, such as in example (27)—a usage that continued far into the 19th century—the verb features the woman in a formally 'intransitive' construction omitting the fetus from the scene altogether. The crucial question is: what is this construction? It does not fit into the ergative system, there is no such thing as an objectless ergative (i.e. an IS-PROC constellation). Is it then to be regarded as a one-participant instantiation of the transitive paradigm? If so, there are two possibilities: either it is an intransitive construction like John died or it is an objectless transitive like soldiers trained to kill. At this point, it is impossible to provide conclusive evidence one way or the other. The easiest way out would be to stipulate that the usage has entrenched 72

Both verbs also occur in literal usage; I have selected these metaphors on the grounds of their corresponding date of attestation. For the sake of clarity, I chose to leave the idiosyncratic abortivate ("to render abortive") out of consideration, since it blends in well with the rest.


203

as a semantic and syntactic translation of the originally intransitive Latin process. The intransitive nature is, however, restricted to literal usage; all metaphorical usages attested occur in an effective construction, e.g. example (34) above. One could argue that the verbs abortive or abortivate were used if an effective construction in literal usage was needed, but supporting data are lacking. In earlier work (e.g. Lemmens 1997), I argued that the construction should be seen as an unusual objectless ergative,explaining the exceptional focus on the Instigator on grounds that this Instigator, the woman, was seriously affected by the event too. In the face of the absence of an ergative effective in literal use in the OED as well as some other concordances of 14th to 17th century English writings, I have come to a more careful stance, realizing that an entrenchm of a 'true' intransitive construction for this particular usage may not be excluded altogether. Nevertheless, some elements against the intransitive interpretation are to be considered. First, in Medieval Latin the abort-verbs (aboriri, abortin, and abortare) could occur in a two-participant construction (cf. example (18) above), with the woman instantiating the role of the Instigator and the fetus that of the Medium. This pleads against saying that the English intransitive is a carbon copy of the Latin possibilities. Secondly, an element in favour of an ergative interpretation are participial constructions as in (35)

[I] could not tell whether to reioyce to see mine aborted infant revived. (OED, 1604)

which is ambiguous between a non-effective reading ("the infant aborted") or an effective one ("the infant that was aborted"). Such ambiguity is typical of the ergative paradigm, and frequently shows up in metaphorical uses as well. In any case, the status of the 'intransitive' the woman aborted (of a son) remains uncertain at this point. To summarize, the situation for abort in the 16th and 17th century was one of considerable flexibility (perhaps even instability) and may very well have been influenced by Latin, French and Italian, languages to which the English Renaissance scholar readily turned in his quest for knowledge and inspiration. In due course, English crystallized much of this diversity into a more coherent ergative whole. This crystallization fits into the global ergativization process that has been show to affect starve, hunger and the SUFFOCATE verbs as well. By the middle of the 19th century, abort seems to have stabilized somewhat into a clear ergative pattern, represented here by the following sentences from the OED (both from Darwin's work), which conceive of a biological development as either aborting (36) or being aborted (37):

204


(36) (37)

If the discs had been small...we might have concluded that they had begun to abort. (OED, 1862) When this occurs, the adherent nectary is quite aborted. (OED, 1859)

Uses such as these may have superseded the verb abortive for which the OED lists no more citations after 1700. The adjective abortive has been severely pruned in its semantic coverage, being restricted to metaphorical usage (which, as concordances show, has always been the item's most common usage) designating an (ergative) conceptualization of an entity or process failing to materialize or coming to success: (38)

Two slips of ground, half arable, half overrun with an abortive attempt at shrubbery. (OED, 1827)

As illustrated in section 7.2 above, this usage has continued up to the presentday English. The demise of the woman-oriented usage (e.g. abortive women) is probably caused by the overall ergativization of the cluster, which gravitates towards the Medium, especially in metaphorical usage. In the course of the 18th century, the adjective abortive extended metonymically to refer to the imperfect result of a failed development. Such an extension is virtually the same as the one from fetus to child: both feature the end-point of the process regardless of whether or not the development has been successfully completed. Quite naturally, this metonymical relation also occurred with the past participle, especially in adjectival use. Compare: (39) (40)

The thorns of plants are abortive branches. (OED, 1834) Thorns, such as those of the rose, are aborted branches. (OED, 1867)

If thorns are aborted branches, they are the deformed remainders of what once were branches (or might have developed into branches were it not for the circumstances that no longer allow them to), just like a broken vase is no longer a vase but fragments resulting from the breakage. The participial uses, including these metonymies, tie in nicely with the ergative paradigm. Also the noun abortion takes on the metonymical reference to the deformed or imperfect result and, by doing so, supersedes the noun abort(e) that in earlier times had such a metonymical extension: (41)

Lest the Child, for want of Nutriment, prove an Abort (OED, 1671)


(42)

205

The skull of the Common Fowl differs in the abortion of the posterior crus. (OED, 1870)

Triggered by this metonymy, the word was for some time spelt abhorsion (see e.g. (24)), as if related to abhor (Shipley 1945:3). More importantly, however, is that the noun takes on the meaning "a non-spontaneous procuring of delivery with the intent to destroy the offspring", a meaning that nowadays is the item's prototype. (43)

The practice of abortion was one to which few persons attached any deep feeling of condemnation. (OED, 1869)

Whereas in earlier uses of the noun aborte, the non-spontaneity of the event was still coded explicitly (e.g. prouokyng aborsum or makynge aborte in examples (21) and (22) above), it now seems incorporated in the noun itself. The "red herring" in the cluster is still the 'intransitive' construction that refers to a woman having a miscarriage. (44)

A woman who aborted at the sixth month. (OED, 1859)

Its continued use indicates that it was relatively well-entrenched, despite its not being sanctioned by the ergative paradigm. Another interesting indicator of its entrenchment is that around the same time that abort was introduced into English (the first half of the 16th century), the (indigenous) intransitive miscarry also came to refer to a premature delivery of a child. I assume that the increasing semantic proximity of the two items has triggered certain lexical and constructional changes. Let us briefly look into them. The use of miscarry to refer to a premature delivery of a child was an extension from a verb that, since the 14th century, had adopted as its most common meanings "to misbehave" (cf. OF meskarier "to go from the right path"), "to come to harm, misfortune or destruction", "to perish", or "to be destroyed". Similar to the abort-complex, the verb (but also the noun miscarriage) extended to the meaning "to fail in one's purpose or object", a usage first attested in writing at the beginning of the 17th century: (45) (46)

... there is none in which Authors are more apt to miscarry than in Works of Humour (OED, 1711) ... you will miscarry if you describe him as a bull. (OED, 1875)

This meaning found a common extension to the legal domain, especially in nominal usage e.g. α miscarriage of justice. As far as I can tell, this phrase is the only relic of this meaning in contemporary English.

206


Most likely, the reference to failing in one's purpose has laid the basis for the extension to the domain of child birth, when a woman delivers before her time and thus fails to achieve her (temporary) purpose. In its reference to a premature termination of a pregnancy, miscarry took the woman as figure (although there are two isolated attestations in the OED that feature the fetus, but these may well have been idiosyncratic as both are from Shakespeare's Henry IV.2, 1597). In short, from the 16th to the 19th century, the two verbs encroached on each other's conceptual space and competed both on the metaphoric and the literal level, with the important difference that each verb had its prototype tied to different domains. In their reference to a woman who delivers prematurely, abort and miscarry share an interesting feature which deserves some attention here. Until some time in the 19th century, miscany invariably had an ofconstituent with the fetus or child as its prepositional object (highlighted in italics): (47) (48) (49)

She had dyverse tymes mis-caried of chylde (OED, 1560) She had once miscarried of a child (OED, 1715) ... a lady ... who had miscarried of a fetus under three months old(OED, 1822-34)

The of-constituent functions to specify what kind of objective has not been achieved, a straightforward extension from the preposition's etymologically basic meaning, "away (from)". This meaning is related to many uses of the particle off as already illustrated in the previous chapter, which derived from the same source. The explicit surfacing of the fetus in the of-constituent is not unwarranted as it functions precisely to identify in what respect the woman has been unsuccessful. It is interesting to see that of also appears in later more general usages of "being unsuccessful": (50) (51)

We dare not risk [our talents] into public view, Lest they miscarry of what seems their due (OED, 1781) He has... pitiably miscarried ofhis object (OED, 1832).

The rather late attestation of the of-complement in this meaning may indicate that the meaning "deliver a child prematurely" was gaining ground as the new prototype, and the other uses (such as these two) required an explicit reference to the unreached goal. Given the common conceptual ground of miscarry and abort, the occurrence of an of-constituent with the verb miscarry may have led conceptually related abort to allow for an of-constituent as well. I should, however, point out an important difference between the two. When miscarry came to refer to a


207

premature delivery, the of-construction introduced a "child" or an "infant", which designated the end-point of the process, the missed goal. In case of abort, however, the of-construction invariably adds a specific characteristic of the fetus (or the infant). Strikingly, the later OED attestations of miscarry with an ofconstruction also contain further specifications about the child (e.g. her third child, a fetus under three months old). This may be taken as evidence of miscarry evolving towards a new prototype in which the fetus is a salient part of the conceptual base: an overt specification in an of-complement will only surface when more than the default reading is at issue. However, without a quantitative analysis of a representative number of attestations for both verbs— which could determine the onomasiological salience of the respective conceptual clusters73— the developments sketched so far remain somewhat hypothetical, despite their plausibility. Legend abort a1 : spontaneous abortion a2: metaphorical extensions a3: non-spontaneous abortion miscarry m1: misbehave, fail to achieve purpose m2: "miscarry" of chylde

lexical meaning construal type

Fig. 23: Abort vs. miscarry in 16th to 20th century English It is difficult to say which factors led to the present-day distribution of the two verbs; in metaphorical terms, this distribution presents itself as the outcome 73

Simplistically put, this indicator of onomasiological salience is the ratio with which an event is coded as a miscarriage or as an abortion. See, for instance, Lemmens (1993) and Geeraerts et al. (1994) for applications of the measurement of onomasiological salience.

208


of a retaliation game. Figure 23 is an attempt to capture its main stages74. For miscarry, the meaning "to deliver a child prematurely" (m2) gradually became the new prototype and the verb eventually lost, as far as I can see, all its other meanings (conveniently summarized as m1). Abort yielded its 'intransitive' "to have a premature delivery" to miscarry, but maintained, at the expense of its competitor, its metaphorical uses (a2) As an instantiation of the ergative paradigm it was well-positioned to continue these Medium-oriented uses. Stage III presents the contemporary situation, which will be elaborated on in section 7.4 below. As emphasized before, the above historical description is far from complete. I have not investigated an extensive, representative corpus for the different time periods nor have I paid much attention to how (and to what extent) English may have been influenced by languages other than Latin (notably, French and Italian). Despite these shortcomings, the main tendencies in the diachronic evolution of the abort-complex can be summarized as follows. The two waves of borrowing abort-forms from Latin—a language that in the time between the first and the second wave had been subjected to a figure/ground reversal from fetus to woman—led in early Modern English to some instability with regard to participant profiling. Via a process of ergativization, however, English crystallized both construal types with their respective profiling into a coherent conception that envisaged the fetus and the woman (or their metaphorical counterpart) as the Medium and the Instigator respectively. Given that in their literal use, the abort-forms typically referred to a spontaneous abortion, such a conception makes good sense: in this conception, the woman (or anybody else) has little control over either the abort process or the Medium, conceptualized as potentially capable of initiating the abortion. Even in the case of an abortus provocatus, the ergative conception holds since, no matter whether the abortion was provoked or caused, the Instigator remains external to the process. Given that in the previous centuries, medical practice had not yet advanced to controlling a pregnancy, an ergative conception that attributes a high degree of co-participation to the fetus is most meaningful. As a rule, the metaphors also settle for this conception, selecting a type of Medium that can either "abort" or "be aborted". 7.4 Transitivization of abort in Modern English The ergativization process that is said to have been operative in Early Modern English means that for a great number of English verbs the process 74

This type of presentation has been inspired by Geeraerts (1994). Note that in stage II of the diagram, the circles representing m2 and a1 have been shrunk to fit in the overlapping area. It is the thickness of lines, not the size of the circles, that signals prototypicality.


209

need no longer be represented as being controlled by an intentional Agent but can be coded as being self-instigated. In the discussion of other manifestations of this ergativization process (Chapter 4, section 4.3.1), it was observed that for some processes, the advent of modern technology and automation may have triggered the conceptualization of various types of machines and technical devices to have their own energy source. In addition to Davidse's examples of the ergativization of intransitive processes with inanimate Actor (e.g. to run the motor), the ergativization of the originally transitive throttle (down) an engine illustrates such a reinterpretation. Ironically, the advent of modern technology and modern health care in particular have had an opposite effect on abort. Modern science has made it possible for spontaneous abortions to be prevented more efficiently than before, and, more importantly, has also brought about the reverse, viz. an easy access to drugs or clinical facilities which enable women to terminate their pregnancy prematurely, without many risks. These factors have contributed, and still do, to a sharp reduction in the self-instigatable character of the abortion process. More and more, the process is moving away from the ergative paradigm into a linear transitive model, in which the parent (or the doctor acting on behalf of the parent) is seen as doing something to the fetus, a participant that no longer coparticipates in the process, but is merely an affected entity. While the technological changes in Western society may have initiated the move of literal abort into the transitive realm, the ethical discussions will ensure that the transitive pattern almost irreversibly continues to entrench as a new prototype. As is well known, the issue of criminal abortion (note the modifying term) has become an intricate ethical, judicial and political matter, which, especially in the US, has led to many a heated debate between pro-lifers and pro-choicers, and which has touched on the fundamental ethics of modern (medical) technology. As a result, people have stopped using the abort forms to denote a spontaneous termination of a pregnancy and have instead resorted to more neutral terms, such as to miscarry, to have a miscarriage or to lose one 's baby. Undoubtedly, influential in this development has been the evolution of the noun abortion, which in present-day English is essentially restricted to referring to a non-spontaneous abortion. As already mentioned above, it is the only meaning that this noun has in the WSJ corpus. It will also come as no surprise that pro-lifers, for ethical or rhetorical reasons, often level abort on a par with other, more typically transitive codings, such as kill, slaughter, or murder. Such parallels, morally justified or not (see below), will certainly enhance the transitive character of abort. This is not to say that an ergative construal of the literal abort-event is no longer possible, but such usage has become marked and basically restricted to medical jargon, as illustrated by the following sentences.

210


(52) (53)

In case of this type of genetic malformation, the embryoME will abort. In case of this type of genetic malformation, an embryoIs will abort itselfMe.

Such ergative construals refer to a spontaneous abortion and not to an abortus provocatus.75 In medical jargon, the noun abortion can also be used to refer to a miscarriage. Recall from the presentation of the contemporary data (section 7.2) that a one-participant construction is still possible in literal use (examples (10) and (11)). Or consider also the following example: (54)

Too many pregnant teen-agers are urged to take the "easy way" and abort, convinced by twisted logic that it is kinder to abort than to bear the child and place it for adoption. (WSJ)

In these examples, abort has acquired the meaning of deliberately procuring an abortion. In these usages, the subject-verb construction realizes an objectless transitive. The reason for this is twofold. First of ail, in present-day English, an 'intransitive' variant without Affected, as in the Queen aborted o f a son, is no longer possible (the objectless transitive still implies an Affected). Secondly, in the new conception of the abort-event, changed because of scientific advances and coloured by ethical considerations, the 'true victim' of the event is the fetus; it is no longer a Medium that somehow contributes to the process or may instigate it but is degraded to a Goal subject to the control of the parent. The second abort in example (54) may be considered as a nice intermediary, since it occupies the middle ground between a transitive objectless and a transitive effective with postponed object {it is kinder to abort [the child] than to bear the child). It can, however, be reiterated that the transitive objectless may not be equated with those uses that explicitly extend to the right (cf. Chapter 5, section 5.4). For the present purpose it is important to see that the construction is now a full instantiation of the transitive paradigm, whereas formerly its status was undecided (either an unusual Is-PROC or and unusual Acintr-PROC). There may, however, be an alternative account for the one-participant construction, which also brings into the discussion the new formations the 5 I did a small enquiry on the Internet (the global computer net) in which I asked people to judge the acceptability of sentences like (52) and (53). Although the results of that enquiry are statistically unreliable, it is significant to note that some people reject these constructions altogether. Strikingly, when a modal commentary was added, the fetus will {most likely/readily} abort, the acceptability increased, which is another indication of the growing transitivity of literal abort. On the interaction between modality and transitivity, see Rice (1987a,b)


211

doctor aborted the woman and aborted women. When the two constructions, the woman aborted and the doctor aborted the woman, are compared, one realizes that they can both be seen as reinterpreted instantiations of the ergative paradigm. In the conceptions coded by these constructions, the woman has taken on the role of the Medium and the doctor that of the Instigator. This conception is quite plausible in view of the technological advances of medical science, which downgrade the involvement of the fetus in the whole process and bring in the responsibility of the medical staff. The reinterpretation shifts the event episode further to the left on the action chain, which results in a conception in which the doctor instigates a premature delivery, a process which requires physical co-participation of the woman. (A similar line of reasoning can be extended to the Dutch construction cited earlier, see footnote 65.) Although the account is a most plausible one, it has to be stressed that this ergative reinterpretation has not yet stabilized into a generally accepted pattern. Firstly, many people still reject a construction like the doctor aborted the woman (in the Internet enquiry (see footnote 75), an overwhelming majority found it totally unacceptable or very bad at best).76 Moreover, while such an ergative reinterpretation is still feasible for example (54) (and example (10) above—to make sure she has completely aborted), it is not for example (11)— the mother can abort, if she so chooses in which the mother is cast in the role of a volitional Actor who has the right (and can therefore choose) to perform the act. The sentence realizes an objectless transitive, with maximal focus on the Actor. Thirdly, the intransitive construction is a minority (in the WSJ, only 3 out of 11), while a two-participant construction with the woman as Agent and the fetus or child as Affected is more common. These codings do not fit into the new ergativization. I therefore believe that at present the literal uses evoke the transitive model, yet they may, at least for some speakers, gradually move into an ergative conception once more, accompanied by a change of participants. What is interesting is that for two-participant transitive constructions, the choice of object is significant. Mostly, it is a child, a baby, or a human being, although, as said, it may also be the fetus or the pregnancy. In other words, in the typical case the end point of the pregnancy is selected, rather than the developing entity. This metonymy may be determined by the aforementioned ethical considerations as well. More specifically, the preference for a certain 76

Interestingly, the one person that found the usage perfectly acceptable, totally rejected the construction in example (53) the fetus will abort itself which may be indicative of a changing conception of participants. Recently, I saw a poster by an anti-abortionist group with a picture of an unborn baby. The caption above it read (translated from Dutch): "If people tell you abortion is a decision which concerns the woman and her doctor only, they are forgetting someone". This is an overt reaction against the recent tendency to demote the fetus in the conception of abortion.

212


profile may be inspired by the speaker's stance with regard to the issue of abortion: pro-choicers may be tempted to select expressions or terms that downplay the "life" feature (such as fetus, (pre-)embryo, zygote or pregnancy) while pro-lifers may try to emphasize the feature by selecting terms like child, baby, human being. The WSJ corpus as well as a collection of newsgroup postings provide some interesting examples of biased coding decisions. For instance, in a WSJinterview, an anti-abortionist is quoted as saying that abortion is murder, life and death broken bodies. A pro-lifer stresses that the strength of the pro-life movement rests in its putting a stop to 16 years of child-killing. Another writer who also argues against abortion goes even further in the end-point focus: he imposes an egocentric profile onto the fetus saying that there was no freedom of choice for me as to whether I should be aborted or allowed to live. Had there been, I would not have chosen to be aborted (my emph.). Especially in the newsgroup postings, members of the other camp mostly use the word zygote. Significantly, a recent study by Coulson (1992) confirms my analysis. In general, her study, which is totally independent from mine, explores "the interaction of cultural knowledge, rhetorical strategies, and pragmatic aspects of language use in Americans' arguments about the morality of abortion"77. In the last part of her paper, Coulson discusses an excerpt from "The Ron Reagan Show", part of a discussion between talk show host (Ron Reagan, RR) and Joseph Scheidler (JS), the executive director of Pro-Life Action, with an interjection from Randall Terry (RT), director of the anti-abortion activist group "Operation Rescue": RR: JS: RR: RT: RR: JS: RR: JS:

77

OK. Joseph, if one of your daughters is raped and becomes pregnant, should [she] be able to have an abortion? I become a grandfather then. If, if, god forbid, and I'm not suggesting this is a possibility, you committed incest with your daughter and she became pregnant— The professionalism of this show is just going down the crapper. No, it happens. Incest is a reality and girls become pregnant because their fathers impregnate them. Ron, that is the most unnatural thing I can think of. I agree but the, should the result of that unnatural act be expunged in any way? It is so repugnant, but to answer your question, I would never kill a child because of the sin of the father. I just would not do it.

References to Coulson's paper will be without page numbers, since I received a plain ASCII version of the paper via electronic mail on April 27, 1995. Full reference to the paper issue (unavailable to me) is in the References.


213

RR:

That's all I wanted to know. Got to go to a commercial. Be right back. RR: OK, we're back with our discussion. Is there any circumstance, Joseph, in which it is proper to abort a fetus, severe birth defects for instance? JS: No. (from The Ron Reagan Show, "Abortion: The Hidden Agenda"; 08/10/91, quoted in Coulson 1992) Scheidler answers negatively to the acceptability of abortion in case of rape or incest. This accentuates the extremity of his position. For the present purpose, it is most relevant to see that Reagan's and Scheidler's statements appeal to different cultural models which condition the coding. In particular, their statements reflect different conceptions of the ontology of the fetus: "although they share the proposition that abortion eliminates the fetus, their differing cultural models have different inferential properties and correspondingly different moral implications" (Coulson 1992). She diagrams the difference of presuppositions as follows:

fetus abortion

= =

Reagan

ScheidSer

result of unnatural act expunging incest

child killing child

In Reagan's frame, abortion equals expunging the result of an unnatural act which should not have happened in the first place. Scheidler, in contrast, presupposes that all pregnancies represent the beginning of life. The pregnancy in the case of rape or incest is, in his view, the good outcome of a bad event. His answers highlight the existence of the fetus. Moreover, in Reagan's personalized scenario of the rape (with which Reagan hopes to make Scheidler sympathize with the victim), Scheidler sees himself related to this fetus (I become a grandfather then). In his model, to terminate this (hypothetical) pregnancy amounts to killing his own grandchild and is therefore morally unjust. Depersonalized, his statements mean to say that the fetus is a child. Coulson's discussion illustrates that a speaker's ideological stance influences how he reasons and talks. In the case of the Ron Reagan excerpt, the ideologically coloured presuppositions of the interlocutors is quite explicit and they themselves are well aware of the ideas they have to argue for. This is also true for the cited letters to the editor in which pro-lifers condemn abortion. Many of the letters to the editor make statements similar to Scheidler's and most writers explicitly equate conception with the beginning of life.

214


Another type of ideologically determined variation that emerges from the data is the lexical variation when it comes to coding the abortion event. For prolifers, awareness of the presuppositions that any pregnancy is the beginning of life and that the fetus is a child to be born has inspired variant codings in terms of kill, slaughter or murder (recall from Chapter 5, section 5.4, an attestation from the LDC in terms of butcher). The following table highlights the lexical variation attested in the WSJ: LEXICAL ITEM

abort kill take life murder slaughter homicide TOTAL

FREQ

PCT

668 6 3 1 1 1 680

98.2% 0.9% 0.4% 0.1% 0.1% 0.1% 100.0%

The high number of occurrences for the noun abortion makes a statistical analysis of the WSJ data unreliable. If it is excluded, only kill remains as an alternative coding, occurring in 35% of the cases (6 versus 11 attestations with abort). Clearly, the coding decisions may not always be as deliberate as in the examples discussed. Often, a certain type of coding is the exponent of the general Zeitgeist in which speakers participate, usually without realizing it. From a preliminary view, my on-going analysis of newsgroup postings seems to provide further evidence that the coding variations are indeed systematic and representative of the ethical conscience of the larger speech community or groupings within it. Returning to the semantics of the cluster in the WSJ corpus, it should be observed that the metaphorical uses of the abort-cluster have not been subject to the transitivization that shapes the literal uses, but continue as instantiations of the ergative pattern. In metaphorical uses, the process can still be coded as instigatable and the affected participant can be attributed an activity concurrent with that of an external Instigator. Given the impact of the ethical debate on the transitivization, the conservative nature of the metaphorical uses seems entirely natural, since there is no ethical need for these metaphors to be recast into transitive ones: developments other than pregnancies can safely be said to either abort or be aborted. In due course, this may leave a semantic network that is somewhat out of balance: the ergative construal will disappear in the literal use, i.e. when referring to the abortion of a pregnancy (in fact, it has already disappeared to a great extent) but the metaphorical extensions still allow an ergative coding. An ergative non-effective coding is, for example, quite produc-


215

tive in computer manuals where executable programs are often said to abort.78 As is typical of ergative non-effective constructions, it is left open whether the process is internally or externally instigated. As already illustrated before, past participial constructions share with the non-effective constructions this voice syncretism, often leaving the issue of instigation unresolved. For example, an aborted takeoff is ambiguous between a takeoff that aborted and one that has been aborted. In line with earlier analyses, the difference between an effective the pilot aborted the takeoff and the noneffective or participial constructions can be represented as follows:

Fig. 24: Participial construal (a) vs. effective construal (b) with abort Note that the participial construction does not profile time as it designates an atemporal relation, i.e. the resulting state of the process that is its base. In the case of a true non-effective construction {the takeoff aborted) the temporal dimension would be profiled. For the sake of clarity, a simplified diagram of a takeoff process has been added; the dotted lines indicate the relationships of correspondence. Although sometimes the larger context may resolve the question of instigation, the participial construction itself remains ambiguous. Often, however, there is no clear indication of an instigating factor, e.g. (55)

"Love Letters" turns sober in its second half as the consequences of this aborted romance manifest themselves in Andrew's unsatisfying but socially correct marriage and, most tragically, in Melissa's divorce and emotional decline. (WSJ)

The broader context does not specify what halted the romance that Andrew and Melissa developed in their love letters; the instigation is no more than a hint. 78

For example, in the on-line UNIX manual on the Motorola Multi-Personal Computer, 40% of the verb forms are used in an ergative non-effective construction (e.g. in case of flag 0, the routine will abort immediately).

216


Figure 25 tries to visualize the main findings with regard to the meaning of the present-day abort-cluster. While focusing on the verb abort, it is considered sufficiently representative for the semantic structure of the entire cluster. The figure is still a simplification of the actual situation, omitting, for instance, the temporal dimension and the distinction into different states, which are both an essential part of full representation of processes but are taken for granted in the present diagram. The diagram nicely shows how the literal and the metaphoric uses of the verb instantiate the two different models of causality. Within the literal domain, the meaning structures "abort a fetus/baby/child" are indicated in bold as the cluster's prototype. For the sake of completeness, the marked ergative non-effective the fetus aborted has also been included in the diagram (the markedness is indicated by the brackets). Obviously, it is still part of the source domain of abort. In order not to clog up the diagram too much, I have left out the alternative account in which the doctor aborted the woman is considered an INSTIGATOR-PROCESS-MEDIUM constellation. There is a final note to be made on the ergative metaphors of the abortcomplex concerning the reappearance of the noun abort in American English. The Supplement to the OED signals a new use of this noun, which has emerged around the middle of this century and is used in aeronautics, referring to "an abortive flight by an aeroplane, etc. or an aeroplane that so fails", as in The control room had to classify our flight as an air abort. Similar usage has been attested in the aforementioned on-line UNIX manual, e.g. in case of abort or abort flag. Such usages can, despite a great deal of semantic overlap, hardly be considered a continuation of the 17th century noun abort for which the OED list no attestations later than 1671 (cf. example (41) above—Lest the Child...prove an Abort). Rather, it seems appropriate in this particular case, to see the reemergence of the noun as an independent phenomenon, a special case of what Geeraerts (1992:188) calls "semantic polygenesis". Given the domain in which the new use has originated, i.e. aeronautics, it makes good sense to assume that the two uses are diachronically unrelated. Yet, as Geeraerts observes, the phenomenon illustrates "the existence of transient meanings in the diachronic development of lexical categories" (1992:188) and is at the same time an indication of the synchronic flexibility of word meanings. 7.5 Conclusions A number of important findings emerge from the diachronic and synchronic description of the English abort-cluster. In the first place, I have demonstrated that for a coherent account of the semantics of abort, one must also consider the paradigmatically conditioned constructions in which it occurs. The influence of the ergative system has, for instance, encouraged the demise of the woman-


217

Fig. 25: A schematic network of the meanings of abort oriented use of the adjective abortive. Secondly, the interplay between the transitive and ergative paradigms clarifies the constructional preferences of literal and metaphorical abort. Conversely, the semantics of abort influences its constructional potential. The ergative paradigm that governed the abort-cluster in the previous centuries is well-motivated in view of the fact that in that period the item referred to a

218


spontaneous abortion, The conception of such an event involves two central participants, the woman and the fetus, fulfilling the role of Instigator and Medium respectively. The woman is not in control of the process, it is the fetus that is the primary energy source of the process and whose co-participation is crucial, even in the case of an effective construal in which the woman or the metaphorical equivalent) is conceptualized as the initiator. However, the role of the woman in the abortion process is to a certain extent ambiguous, as she is also considerably affected by the process. This sanctions an unusual 'intransitive' construction, which can either be regarded as a true intransitive with an unvolitional Actor or an unusual 'objectless ergative'. Despite the separate status of this construction (on either interpretation), it appears to have been well-entrenched. More extensive data-analysis may provide more insights into its paradigmatic niche. That experiential factors influence the conception of abortion is irrefutably demonstrated by the transitivization of abort in contemporary English. Against the background of the medical and technical advances through which people have gained (more) control over human physiology, the conception of abortion has shifted to a transitive one in which the woman controls the process. In the extreme case, she herself has the right to decide whether or not to terminate her pregnancy. In this conception, the fetus is no longer seen as potentially initiating the process, but figures as an inert Affected. The transitivization is further enforced by ethical considerations: to avoid the difficult issue of abortion, people will, contrary to what used to be the case, refrain from using any of the abort-forms to refer to a spontaneous abortion. Anti-abortionists take the transitivization one step further by equating an abortion with the killing of a human being. My data as well as those presented in Coulson (1992) indicate that the rhetoric of anti-abortionists is inclined to profile the child as the Goal of the transitive abort-process. So, the transitivization accounts for both the constructional and the collocational tendencies in literal usage of abort. This process of transitivization has led to a peculiar imbalance in the semantic network of the abort-cluster while instantiations referring to a premature termination of a pregnancy are governed by the transitive system, the metaphoric uses continue to realize the older ergative paradigm. The paradigmatic contrast deepens the division between the literal and metaphorical uses. Undeniably, the two are still semanticaily related—one would not want to characterize the abort-cluster as homonymous—but the degree of similarity is certainly reduced because of the additional paradigmatic breakpoint. Note further that the transitivization of literal abort does not undermine the ergativization that has characterized English over the last centuries. It only shows that, at least for literal abort, the ergative construals have become quite unconventional, since they are no longer compatible with the verb's meaning.


219

Against the background of the action chain model, the transitivization of abort can be thought of as one move up against the flow of energy, as a result of which the woman becomes the most central participant and the fetus (the downstream participant) is demoted. Sentences like The doctor aborted the woman or the woman was aborted shift the profile on the action chain further to the left, including in the event schema the doctor as the agent who is ultimately responsible for the abortion. For some speakers, the woman may even take on the role of the Medium, seeing her as actively participating in the process; the sentence the woman aborted is then no longer interpreted as an objectless transitive but as an ergative non-effective. However, at present, many people still strongly object to the construction the doctor aborted the woman. It is an open question whether this re-ergativization will establish itself more firmly.

CHAPTER 8 SUMMARY AND EVALUATION By looking in more detail at the semantic structure of a number of verbs from the field of killing, the present book has substantiated as well as refined the view that the English grammar of causative constructions is governed by the transitive and the ergative paradigms. A major refinement of Davidse's model pertains to the weight given to lexical meaning as determining the range of constructions in which a verb can occur. As such, the present book situates itself in between the purely syntax-based and the purely lexically oriented approaches. This work has shown that the meaning of a verb may bar certain constructions that from a purely grammatical point of view should be allowed; on other hand, it has shown that constructional shifts may come to expand the lexical structure of the verb as well. Before broaching a more critical discussion of the interaction of the lexical and constructional meaning, the main findings are briefly recapitulated. 8.1 Summary In a critical evaluation of Levin (1993), a different categorization of verbs of killing has been proposed, predominantly based on the lexical structure of these verbs. The constructional properties can be seen as an additional layer of complexity, bringing in the distinction between the transitive and ergative paradigms. For some subgroups, the transitive/ergative distinction aligns with the lexical divisions, as with the MURDER verbs, the DECAPITATE verbs, and the INSTRUMENT verbs which are all prototypically transitive and the STARVE verbs which are all prototypically ergative. Another large subgroup, the SUFFOCATE verbs, contains ergative as well as transitive members, although the group is argued to have a strong predilection for ergativity, which has been shown to be experientially motivated. Moreover, another experiential distinction, viz. that between internal and external causation, was argued to underlie firstly, the distinction between the ergative and transitive SUFFOCATE verbs, and secondly, the occurrence of the non-effective versus effective construction for the ergative verbs within this group. Obviously, these distinctions are not to be regarded as absolutely predictable but as tendencies strongly motivated by our experience. This book has offered more elaborate descriptions of representative subgroups of the field, set up to cogently illustrate the semantic difference between the two

222

LEXICAL

PERSPECTIVES ON TRANSITIVITY AND ERGATIVITY

paradigms and to elucidate the dynamic interplay between them, both from a synchronic and diachronic perspective. The transitive paradigm, which prototypically involves a maximally asymmetrical relation with a volitional Actor acting on an inert Goal, has been illustrated by the MURDER verbs. Within this subgroup, the transitive effective structures were shown to have a strong implication of goal-achievement, which suggests a refinement of Davidse's view. She sees volitional goal-directed action as the prototype of the transitive effective (see Figure 13 in Chapter 2). For my data, I argue that structures which evoke both goal-direction and goal-achievement are prototypical. In the class of the MURDER verbs, which can be seen as prime examples of transitive verbs of killing, a focus limited to goal-direction (and exclusion of the goal-achievement) has only been attested for the more general verb kill, as in my feet are killing me. The more central members of the MURDER class combine the concept of goal-direction with that of goalachievement. The lower salience of goal-achievement in the semantics of kill explains why it occurs in the objectless transitive more frequently than the other MURDER verbs. While I have rejected a purely lexical account of the objectless construction, I will not deny that the construction may be more common with what Rice (1988) has called "neutral" verbs, although further empirical research is warranted. In the case of kill, the neutrality encouraging the objectless transitive concerns primarily goal-achievement. As a refinement to Davidse's representation, the schematic network of the MURDER verbs can thus be represented as follows:

Fig. 26: Simplified schematic network of the transitive effective The thickness of the lines indicates the degree of prototypicality of the instantiations, at least as it emerges from the data on the MURDER verbs. Since the

SUMMARY AND EVALUATION

223

middle construction is quite atypical for verbs of killing, its box has a dashed line and the distance of elaboration has also been increased.79 Both the middle and the objectless transitive are categorized as effective constructions because the former still implies an Actor, and the latter, a Goal. The investigation has also unequivocally shown that, as a rule, accidental goal-achievement does not occur with the MURDER verbs except, once again, with kill. The reason for the absence of such a construal is that the other MURDER verbs saliently evoke the notion of an intentional Actor who acts deliberately on the Goal. Consequently, sentences like Oswald accidentally assassinated JFK or Terrorists murdered Higgins, but didn't intend to are infelicitous, except when they refer to situations in which the wrong person has been killed. However, even in these cases the Actor has acted volitionally. This, I argued, confirms the claim that volitionality applies to the relation of Agent (Actor or Instigator) to the process and is not some inherent feature of the Agent as such. While the accidental killing of the wrong person is possible with the MURDER verbs, the prominence of premeditation in their semantics renders such construals quite marked. Since the notion of volition is less salient with kill, this verb occurs more naturally in constructions that express accidental goal-achievement. For the latter type, the verb is often attested in agentless passive constructions referring to deaths in accidents. I argued that in such agentless passives the victim is presented as merely inert Affected and the inquiry into the Agent is irrelevant or odd. In related codings with die, the victim is cast in an Actor role, an involuntarily Agent but an Agent nonetheless. English is markedly different from other languages in its common usage of kill in reference to such accidental deaths. Dutch, for example, almost invariably codes such event in terms of an intransitive process like sterven ("die") or omkomen ("die"). An agentless passive of transitive doden ("kill") is restricted to casualties in violent acts (e.g. assaults, acts of warfare or (occasionally) a bombing) which can always be traced to an ultimate (yet implied) Agent. Given its less stringent ties to volitionality, kill was also shown to frequently occur with inanimate Agents, constructions which are not, as is often assumed in the linguistic literature, unusual or marked. My data have revealed that such codings are experientiaily well-motivated. The lexical and constructional behaviour of the predominantly ergative SUFFOCATE verbs have undeniably revealed the Medium-oriented character of the ergative paradigm. As said, the ergative predilection is well-motivated in 79

Recall that instantiation automatically implies schematization. Arrows marking schematization and bi-directional extension arrows between 'sister-nodes' have been omitted here so as not to clog the figure too much.

224


view of the activity of the Affected: in addition to be affected it saliently coparticipates in the conceived process. Such conceptualization has a strong link with our experience of suffocation in the real world: it is the victim that is 'doing' the suffocation, whether or not this process has been externally instigated. The suffocation itself is typically conceptualized as relatively independent of the Instigation, and the cause need not be concurrent with the effect. Undoubtedly encouraging a high degree of independence of the caused process is the typical imperceptibility of the cause, especially those causes typically involving an internal perspective (e.g. gases, lack of oxygen, etc.). The transitive SUFFOCATE verbs, strangle, throttle and smother clearly have a more outspokenly external cause, typically the construction of the throat or the covering of the external parts of the respiratory system. A transitive conceptualization focusing on the Agent inflicted harm on the inert Affected is wellmotivated here, as the affected parts are those which can be easily accessed by an external Agent. The external parts are also those that usually are not conceived as active within the respiration system. As is typical of transitives, strangle, throttle and smother altogether omit the notion of the victim's participation. In contrast to the MURDER verbs, the SUFFOCATE verbs show considerable lexical and constructional flexibility, both synchronically and diachronically. In diachronic perspective, the flexibility takes the form of a relatively fast evolving convergence in which the different verbs display a high degree of lexical and constructional interchangeability. This lexical convergence can surely be attributed to the experiential proximity of different 'types' of suffocation events (such as caused by constriction, coverage, poisonous gaseous, etc.) as well as the variability in the conceptualization of a cause. The constructional convergence lies in the overall ergativization of the verbs spanning a period of 500 years or so. As such, they can be taken as substantiating Halliday's claim that such a process of ergativization has affected Modern English to a great extent. Whether this is true for verbs from other lexical fields and, if so, what motivates it for these verbs, requires further empirical research (cf. below). In the case of the SUFFOCATE verbs the ergativization process is, as said, experientially wellmotivated. A synchronic perspective reveals that the SUFFOCATE verbs still display a certain degree of overlapping, especially in their sharing similar metaphorical extensions, such as the suppression of speech or their frequent hyperbolical usages. However, in contrast to what used to be the case, the SUFFOCATE verbs have moved to more differentiated conceptual centres, mostly a reorientation towards their etymologically basic prototype. As shown, these different prototypes are reflected in different collocational patterns, in literal and metaphorical uses. The more differentiated conceptual centres have also led to a


225

more strict constructional divergence. In present-day English strangle, throttle, and smother behave as full-fledged transitive verbs, although the latter was attested in what can be regarded as a continuation of its earlier ergative uses. In present-day English, there also appears to be a stronger association between a given meaning of a verb and a given construction. A case in point is for example the conceptualization of swallowing the wrong way or that of an accidental drowning which either exclusively or prototypically occur in a non-effective construction. Such coding is well-motivated, once again, in view of the salience of the activity of the Affected. Significantly, when the typically ergative verbs choke, drown and suffocate are taken together (and the phrasal verbs excluded), the preference for a noneffective construction is overwhelming: it occurs in 63% of the cases. Past participle constructions, which are ambiguous between an effective and a noneffective reading represent 9.5%. In short, in almost 73% of the. cases, a noneffective reading is supported. Yet when these verbs are used to encode a suffocation event caused by a markedly external cause, an effective construction is most typical, e.g. this necktie is choking me. Also, when the cause itself is more active (e.g. a moving entity) or furnished with a higher degree of responsibility, it tends to surface as the Instigator in an effective construction, e.g. the waves drowned me or the ashes entered my throat and nearly choked me. Typical of the SUFFOCATE verbs their hyperbolical use in reference to experiences conceived of as resembling suffocation. Such hyperboles are extremely common with the ergative SUFFOCATE verbs. Think only of expressions like drowning in paper work, a suffocating room, or a choking collar. Not only do these uses exaggerate the experience comparing it with suffocation they also exclude from the conceptualization the lethal outcome of the suffocation event; hence the typical occurrence of an -ing form. The commonality of such hyperbolical uses, which tend not to occur with the MURDER verbs except for kill, can be attributed to the inherent focus on the activity of the Affected, unfolding over a certain time period. Similar comments can be made concerning prototypically ergative starve where the extended duration of the starvation process is very much part and parcel of the verb's conceptual structure. To sum up, the SUFFOCATE verbs show a high degree of lexical and constructional flexibility which finds a natural explanation in view of how suffocation events are experienced. At the same time, the constructional flexibility can largely be explained in view of the semantic structure of the verbs themselves. As such, the analysis of the SUFFOCATE verbs presents an investigation into the complex and dynamic interaction between verbal and constructional semantics.

226


Another example of the experiential grounding of the interplay between lexical and constructional meaning is the polysemic complex abort. The diachronic description of abort has demonstrated how Modern English integrates the two seemingly conflicting patterns from Latin, viz. a fetus-oriented and a womanoriented one, into a coherent ergative system. In this new system, the woman takes on the role of Instigator and the fetus that of Medium. The ergative construal is conditioned by the item's prototypical reference to a spontaneous abortion in which the fetus is seen as co-participating in the event and capable of self-instigating it. In contemporary English, the verb prototypically refers to an abortus provocatus, with the fetus no longer conceptualized as co-participant in the process but as a wholly affected, inert Goal. The ensuing ethical debate has strengthened the transitivization and abort is now often put on a par with verbs like kill, murder, and even slaughter or butcher. In tandem with this conceptual realignment of abort, the rhetoric of anti-abortionists often emphasizes the life aspect of the fetus, selecting terms such as child or human being to refer to the Goal. The transitivization has not affected the metaphorical uses, which therefore continue the ergative pattern. In these uses, the process can still be coded as instigatable and the affected participant co-participates in this process. This strong paradigmatic breakpoint increases the semantic distance between the literal and metaphorical uses. However, this situation is clearly not permanent, as the verb shows to be a fine illustration of 'change in progress'. For some speakers, a new conceptualization is emerging for abort involving different participants (the doctor and the woman) and an ergativization with the woman as the Medium. Many people object to this use, however. Personal comments by various native speakers have moreover shown another possible evolution, viz. a continuation of the transitivization. For these speakers a transitive conceptualization is the sole possibility for metaphorical uses as well. Notwithstanding the common use of a non-effective, they disapprove of constructions like the mission aborted, observing that an effective should be used (they often suggest an agentless passive). Other empirical explorations in this book have offered further insights into various constructional patterns as being influenced by verbal meaning. First, important insights come from a diachronic analysis of paradigmatic changes, other than the aforementioned (Early) Modern English ergativization of the SUFFOCATE verbs and of abort. Secondly, regarding present-day English, it has been shown how the middle construction is constrained by the lexical structure of verbs, how the paradigmatic opposition proves to be relevant for -er derivation, and how the objectless transitive should not be explained on purely


227

lexical grounds, as Rice (1988) does, but against the background of the semantics of the construction itself. In the diachronic analyses, the ergativization of originally intransitive starve stands out as a marked paradigmatic shift. Triggered by verb's lexical extension to a non-volitional process of starvation, the notion of instigatability become a logical coding option. While initially the ergative construal may have been experienced as an extension from the intransitive prototype (much as is the case with present-day constructions like walk the dog ox jump the horse), it rapidly entrenched to become the verb's new constructional prototype harmonizing well with the new lexical prototype. Although one would expect such an entrenchment to occur with the related verb hunger as well, this did not happen, for reasons that we have not yet been able to clarify. Possibly, the entrenchment of hunger as coding an intransitive process of longing was triggered by the semantic proximity of the two verbs; such proximity almost invariably gives rise to semantic change. Further corpus research may shed more light on the issue. How paradigmatic changes can lead to mixed patterns for one and the same verb (much like with abort) has been illustrated by the case of throttle, for which an ergative coding has become possible in the extension to controlling the fuel influx in engines. Our contemporary corpora indicate that the verb is still used as a SUFFOCATE verb, yet many speakers seem to associate the verb exclusively with its usage in the domain of engines and motors. If entrenchment of this usage continues, the result may be a full prototype shift, in which the item's original meaning 'to strangle" becomes a peripheral extension. It is not unthinkable that in due course, the latter may become homonymic with the new prototype, but at this time it is too early to draw definite conclusions on this issue. Turning then to the present-day situation, my investigation of the middle construction has provided some original insights. First, I have demonstrated that ergative as well as transitive verbs can occur in a middle construction. The meaning of this construction focuses on the properties of the Affected which facilitate or hinder the execution of the process as carried out by some implied Agent. The ergative middle is a less typical middle construction because it resembles a regular non-effective construction both in form and meaning. Drawing on constructions with typically ergative verbs of killing, such as starve or choke, I have furthermore shown that while a middle-like construction (e.g. Asthmatic patients choke easily) is possible, it will more likely be interpreted as an ergative non-effective. It is essential to realize that even if this sentence were interpreted as a middle construction (i.e. with an implied Instigator), the notion

228


of co-participation remains part and parcel of the construal, as with regular ergative effective constructions. The transitive middle construction, on the other hand, is more conspicuous as it more overtly departs from canonical transitive syntax: the Goal becomes the subject of the active transitive verb. As already emphasized, despite the active voice syntax, the semantics is that of a passive construction in which the Goal is an inert participant wholly affected by the action of the implied Actor. It does not become an agentive participant. The non-agentive nature of the Goal in middle constructions is confirmed by ambiguous constructions. For example, a killing bull can either receive an active interpretation (a bull that kills) or a passive one (a bull predestined to be killed). This ambiguity is different from that with ergative middle constructions, which reveals once more the polar opposition of the paradigms: the transitive middle may alternate between an Actor/Goal reading, the ergative between a Medium [+sslf_instigatbg]/Medium[-selfinstigating]

one

My research reveals that verbs of killing generally do not occur in a middle construction. Contrary to what Smith (1978) and Rapoport (1993) claim, the explanation for this does not hinge on the degree of affectedness or the notion of change of state. Rather, the semantics of the verb must be compatible with the meaning of the middle construction. With verbs of killing, it is difficult to conceive of an entity as possessing qualities that make it suitable for killing. However, in the right context, even these verbs may occur in a middle (or middle-based) construction. The few attestations of middle based constructions with verbs of killing (e.g. slaughtering cattle) support my position. I have also shown that the occurrence of a verb in a middle construction correlates positively with its potential for an -able derivation. The reason, of course, is that both the middle construction and the -able derivation focus on the properties of the Affected that play a role in the given process. Hence, in this book verbs of killing have been identified as non-middable as they, as a rule, do not allow the middle construction or the -able derivation. Once again, it is possible to design a context in which the latter may occur. For example, a cruel tyrant may be said to be justifiably slayable. Both the middle construction and -able derivation remain, however, atypical with verbs of killing. The unavoidable more general conclusion is that while prototypical cases can be discerned, there are no hard-and-fast criteria for the exceptability of either the middle construction or the -able derivation. Verbs usually not allowing these constructions can in certain contexts be extended to occur in these constructions leading to unexpected construals that are, however, semantically well-motivated.


229

The agentive derivation is another area where paradigmatic influences show up. Within the field of killing, the -er deviation is only attested for the transitive verbs and its transitive preference has further been substantiated by an extensive survey of virtually all -er derivations in the WSJ corpus. At the same time, the description of-er derivations has demonstrated that the agentive nominalization reflects a number of characteristics of the transitive and ergative domains. ' First, my investigation evidenced that the conceptual centre of the suffix is the nominal equivalent of that of the transitive effective, as it prototypically denotes the intentional or habitual Agent of the process designated by the base verb (e.g. (serial) Jailer or strangler). From this prototype, the suffix has extended to profile instruments as in copkiller (bullet) and events, as in this show is a killer. Secondly, contrary to the claims made by Keyser & Roeper (1984), I showed that the -er suffix can profile the Medium of an ergative base process as in e.g. starver or drowner. As is the case with starver, the Medium-oriented -er is often derived from originally intransitive verbs which have become ergativized and which have retained a strong focus on the original intransitive Actor. Finally, quite remote from the truly agentive prototype are the -er derivations based on the middle construction, since they profile the inert Goal of the transitive base process. Contrary to Ryder's (1991b) characterization of these formations as expressing 'agentive patients', I have argued that, like the middle construction, these formations include in their semantics a passive voice constellation: they designate entities predestined "to be V-ed". Significantly, the middle-based formation confirms my characterization of middable verbs, as verbs of killing are shown to strongly resist such middle-based derivations, although also here their prototypical non-middable character can be overruled. Recall, for example, the use of killer in reference to cattle ready for slaughter, an unusual usage of an item that prototypically refers to the Agent of a killing process. My analysis of the objectless construction has argued that lexical considerations, as offered in Rice (1988), tell only part of the story. Instead, I argued that one should consider the construction's semantic import, viz. the maximal focus on the ACTOR-PROCESS unit, mostly implying of a general statement which abstracts away from individual instantiations. Against the background of this non-individuating nature of the construction, the schematicity of the Goal can be accounted for. Clearly, certain elements in the lexical structure of individual verbs, e.g. their being neither extremely schematic nor extremely specific, may be important and be held responsible why, for example, the objectless construction prototypically occurs with these verbs. The high frequency of objectless constructions with kill but its non-occurrence with the other MURDER

230


verbs can be taken as indicative of this fact, but more extensive corpus-based research with verbs from different lexical fields is warranted. A final element of 'unexpected' clause semantics is the conceptual integration of verb and particle with phrasal verbs. My analysis clearly indicates, as is now more widely accepted certainly within Cognitive Grammar, that particles do carry meaning in and by themselves. An important aspect of their semantic contribution to the clause is an increased force-dynamic component, which can be attributed to their mostly spatial prototype. The increased force-dynamic component can be such that the resulting construal has become fully transitive, even when a prototypical ergative verb is used, as illustrated by phrasal verbs such as drown out, choke back or choke down. These construals no longer activate the ergative model, as the Affected entity no longer co-participates in the event but is conceptualized as inert Goal fully affected by the Actor's saliently forceful action. Often also, a change of participants is involved, as in the case of choke down or choke back. Clearly, the importance of particles to clause construal requires more research than we could afford to do within the confines of this book. First, the addition of a particle need not always imply a paradigmatic shift, as was illustrated with choke up. Secondly, it has been noted that when transitive verbs, e.g. kill combine with a particle like off there also seems to be an increase in force-dynamics as it comes to denote total annihilation. It would be worthwhile to evaluate the particle's impact on the constructional properties of transitive verbs in general. Moreover, lexical matters should not be overlooked. For example, kill off is acceptable, but murder off or execute off are not or are awkward at best. Further empirical analysis may shed more light on global patterns and tendencies for the semantic merger of verb and particle. In short, the various descriptions in this book have fleshed out the semantic differences between the transitive and ergative paradigm. The former is Actorcentred, right-oriented and affects an inert goal. The latter gravitates towards the Medium, which in addition to being affected co-participates in the process; often, an Instigator is added to this conceptual nucleus. At the same time, by drawing attention to the lexical flexibility of verbs as influencing their constructional possibilities, the various descriptions have equally shown that a verb cannot be absolutely assigned to either paradigm given that some lexical extensions may sanction constructions that depart from the prototypical constructional specifications. Various such paradigmatic shifts have been observed in the diachronic analyses, showing how such 'extensions' may lead to either full or partial change of a verb's 'paradigmatic address'.


231

Such paradigmatic changes, often triggered from the lexical end, provoke thought on the interaction between lexical and constructional meaning. How are we to understand this interaction? Is assuming such interaction not a case of circular reasoning? More specifically, to determine the semantics of a verb one looks at the constructions in which it can occur, yet at the same time, one characterizes the semantics of the construction on the basis of classes of verbs that can occur in it. Is this really an example of circular reasoning and, if so, can it at all be avoided? A second area of critical thought concerns the distinctiveness of the transitive and ergative constructions themselves. In some usages, a verb may no longer allow alternations that are otherwise prototypical for that particular verb. For example, in some uses, an ergative effective (e.g. The bell drowned out the protests) no longer finds a non-effective construction, an ergative non-effective construction (e.g. Polly choked on a fish bone) may not sanction an effective counterpart. Does this not signal another type of circularity related to the definition of the paradigmatic contrast, since one construction (e.g. the ergative effective) is said to realize a given paradigm on the basis of the existence of certain alternations for that construction (e.g. the non-effective). In the face of the evidence, can the transitive and ergative paradigms still be maintained as being truly distinct models. Should the general grammatical distinctions not be dismissed in favour of a general transitive/intransitive distinction lined up with the over nominative/accusative case marking in English? Critical questions such as those raised here surely do not find a simple answer. This book should be regarded as only the onset of a formidably complex analysis of how the ergative and transitive paradigms do indeed shape the grammar of causative constructions, provided that one realizes how lexical and constructional meaning interact in specific contextualized instantiations. As such, this work does not claim any definite answers to the questions raised above, yet against the background of the analyses presented, some suggestions can be offered. The following section will first take up the issue of lexical and constructional interaction and then move on to an evaluation of the cognitive reality of the transitive and ergative paradigms. As a logical conclusion to these explorations, the final part considers some avenues for further research. 8.2 Evaluation 8.2.1 Lexical and constructional interdependency The meaning of any lexical item involves an imposition of a particular profile on some conceptual domain. In the case of verbs, and certainly verbs of killing, event structure constitutes the relevant domain. What the verbs 'pick out' from this structure includes the nature of event participants and their interaction.

232


Constructions could then be characterized as tying down these participants in an explicit configuration. Focusing on the strong correlation between the meaning of verbs and their constructional range, Levin (1993) categorizes verbs into different classes whose constructional properties are fully predicted by the verbs' semantics. I have argued that such an exclusively lexical-semantic perspective is inadequate. It neglects the meaning of the constructions, which can and should be characterized in its own terms. The constructions themselves constitute semantic elements. Once one realizes that constructions and verbs carry meaning independently of each other, one faces the problem of how the two combine to form a composite structure. One obvious prerequisite is semantic compatibility. The participant roles expressed by the verb and the argument roles associated with a construction can only be 'fused' when they are semantically compatible. Moreover, as Goldberg (1995:50) observes, each participant role that is lexically profiled must be fused with a profiled argument role of the construction. For example, the ergative non-effective construction requires that its profiled participant be an entity capable of self-instigating the process. In the contemporary conception of an abort-event, the fetus does not possess these qualities and consequently, it cannot be coded by an ergative construction. Very often, paradigmatically opposed verbs, which otherwise appear synonymous, will have different types of participants associated with them, a point to which I return shortly. Since constructions carry meaning, it is entirely logical that the meaning of a verb is altered when it occurs in different constructions; any difference in form always signals a difference in meaning. Another logical consequence is that the semantic differences between various constructions with the same verb may be attributable to the construction rather than to the verb in and by itself. Differently put, a verb's prototypical lexical coverage may be extended under influence of the construction in which it is used. For example, the middle construction offers an atypical perspective when used with verbs of killing, yet no doubt it is the Goal-focusing feature of the construction and not some component of the verbs' semantics which 'triggers' the semantic change. In other words, it is the construction that stretches the verb's meaning beyond its typical usage. Similarly, it is the general focus on the action-component characterizing the transitive objectless that takes the verb to what could be called a more 'neutral' use, rather than claiming the reverse as Rice (1988) does. On the other hand, lexical semantic changes may influence a verb's prototypical constructional possibilities. This has been amply illustrated by the diachronic evolution of starve or some of the metaphorical extensions of throttle and smother. When via a series of extension paths throttle comes to refer to


233

regulating the influx of fuel in an engine, the new semantic structure of the verb becomes compatible with the ergative construction simply because the engine can be conceived of as having the possibility of self-instigating the process. Similarly, when starve came to denote a lingering death from hunger or cold, the process could readily be conceived as an instigatable one. In short, these are constructional shifts 'triggered' from the lexical end. Such shifting is also quite prominent in the case of abort: in the contemporary conception of an abortus provocatus the Affected has become an inert Goal which is not seen as capable of instigating the process. In all these cases, the paradigmatic breakpoint loosens the semantic relatedness of the different usages of the verb. As a result, these usages may entrench as separate verb senses that each project a different set of constructions. To summarize, while the meanings of a verb and the construction in which it occurs can be characterised independent of each other, a specific usage will fuse them into a composite structure in which they become interdependent. In some cases, it may be possible to discern whether the meaning of the construction or that of the verb has triggered differences in construal. Yet in most cases, the picture is not as clear-cut and the semantic modulation is a dynamic interplay of lexical and constructional, meaning, in which it is no longer clear what triggered what. A good example is the partial ergativization of transitive verbs like drive, sell, or wash via the middle construction. This ergativization is an instance of ongoing linguistic change in which the semantic structure of the verbs may gradually change towards full compatibility with the semantics of the ergative paradigm. At the same time, the semantic structure of the verbs is such that it allows and possibly reinforces such reinterpretation. Recall that it is precisely the semantic structure of verbs of killing that makes such a middle-triggered ergativization unlikely to occur with these verbs. Constructional and lexical profile are in fact two sides of the same coin. As I see it, asking for an unambiguous answer to the question which side was stamped first is asking the wrong question. It cannot be denied that there is some circularity in this view, yet this is, I believe, unavoidable. We have no other way of discovering verbal and constructional meaning than to examine different usages and abstract from these. Only by exploring a vast amount of individual clauses can one refine insights into the dynamic interdependency between a verb's semantic structure and its constructional range. While initially somewhat circular, this type of analysis takes the form of an ever widening spiral-like exploration into the grammar of English. 8.2.2 The cognitive reality of transitivity and ergativity In specific clause construals a verb may be used in constructions that are no longer prototypically associated with it. When such clause construals lead to

234

LEXICAX PERSPECTIVES ON TRANSITIVITY AND ERGATIVITY

paradigmatic changes, it may be questioned whether the paradigmatic opposition is valid at all. The question is particularly pertinent in the case of the ergative effective construction. This construction, as for example realized by John suffocated/drowned/choked Mary, is said to evoke the ergative model given the possibility of the non-effective, e.g. Mary suffocated/drowned/choked. The recurrence of the same lexeme in the two constructions is taken as evidence that in both case the ergative paradigm is invoked. This cannot be said for The woman aborted the child or John opened a tin of baked beans, as they do not allow the alternation. Thus, lacking a feature typical of the ergative paradigm, the clauses are said to realize the transitive model only. The question could also be applied to an ergative non-effective as in Polly choked on a fishbone which was shown not to tolerate an effective variant, *A fishbone choked Polly. It does not even allow an agentless passive, *Polly was choked on a fish bone (a construction which was shown to be possible with drown). Given that instigatability is one of the defining characteristics of the ergative non-effective, can this sentence still be said to activate the ergative model? Should it not be regarded as a full-fledged non-instigatable intransitive (transitive non-effective)? These questions strike at the heart of the paradigmatic distinctions. Can the transitive and ergative be maintained as distinct models if verbs show considerable paradigmatically 'problematic' instantiations? Differently put, how cognitively real are the paradigmatic distinctions? Some people, especially those uncomfortable with seeing the notion 'ergativity' applied to English, may come to reject the 'ergativity' of an effective clause like John suffocated Mary saying that it is as transitive as John killed Mary. The impossibility of the non-effective in certain uses of prototypically ergative verbs would then merely confirm their point. This book has clearly demonstrated that more careful pronouncements are in order, in which I align myself with Davidse's view yet take a more nuanced position. The underlying assumption is that one should discern two levels on which paradigmatic oppositions manifest themselves: on the phenotypical level, i.e. with overt formal reflections, such as case marking, and on the cryptotypical level, i.e. with more covert categories making themselves felt only indirectly. Ergativity in English belongs to the second group as it is signalled by the systematic use of the same lexeme in the effective and the non-effective. This means that the ergative effective realizes two models: through case marking (i.e. phenotypically) it instantiates the transitive model and cryptotypically it realizes the ergative model. Davidse's concern is primarily with demonstrating the existence of the transitive and ergative paradigm and as such, her description is strongly oriented towards the grammar and less concerned with specific verb-bound, or rather


235

context-bound, constmals. The latter are precisely the main concern of this book and they have revealed the more complex and probably less distinctive character of the two paradigms. Provided one takes into consideration some of the points raised in this book, the view that English is structured by the two paradigms stands up to scrutiny, even for cases like (certain) ergative effectives whose ergativity may not stand out as much. To begin with, and has also been pointed out by Davidse, while prototypically distinct, the two paradigms should be regarded as prototype structured categories certain members of which are situated in the transition area between the two paradigms. For the ergative paradigm, certain effective constructions may realize ergativity only to a small degree or not at all; certain non-effective clauses (e.g. to choke on food) may be situated closer to the intransitive end by not readily allowing an overt Instigator. For the transitive paradigm, certain constructions (e.g. the middle construction) may take some verbs (e.g. sell or wash) to displaying ergative characteristics. Yet I have argued that even for these cases, a more nuanced view is needed, as they do retain the focus on the feasibility original to the middle construction. As a specific case of variability between transitive and ergative instantiations, let us briefly consider the ergative effective. While formally (i.e. phenotypically) identical with the transitive effective, the underlying semantics where shown to be different as the Affected participant, the Medium, co-participates in the process. Is some cases, this independent action of the Medium is confirmed explicitly within the effective, in as Vendler's (1984) example He boiled the egg for half an hour or as shown by anaphoric reference in The government starved the country, but it took them years to bring it about. The possibility of separating the caused process from its instigation signals its independence, is typical of the ergative paradigm. Mostly, however, the co-participation is not so overtly marked in effective constructions. In any case, it depends on the specific usage of the verb, as in some contexts co-participation is no longer involved as shown by the impossibility of a non-effective. It was precisely for failing to note this verb-bound variability that Levin (1993) was criticized, relying too heavily on the general notion of 'change of state'. However, in later work, viz. Levin & Rappaport (1995), the rigid view on categories has been relaxed and they observe that a non-effective may not always be possible, which they illustrate with the following examples: (l)a. He broke his promise/the contract/the world record. (L&R 1995) b. *His promise/The contract/The world record broke. (2)a. This book will open your mind. (L&R 1985) b. *Your mind will open from this book.

236


To account for this, Levin & Rappaport draw on the notion of selectional restrictions, observing that "the set of possible subjects for the intransitive use of a verb appears to be a subset of the set of possible objects for the transitive use of the same verb" (1995:86). In essence, the idea of selectional restrictions is compatible to what has been said above: for the ergative model to be invoked, the slotfiller for the Medium role must posses the possibility of co-participation, i.e. be capable of sustaining and self-instigating the process. In addition to the observations concerning collocational patterns for ergative verbs of killing offered in this work, other research has shown that collocational patterns indeed confirm the notion of coparticipation as underlying an ergative construal. For example, De Coster's (1995) corpus-based analysis of the pair heal/cure reveals that heal, a prototypically ergative verb, typically occurs (also in effective constructions) with a participant that can be conceived of as selfinstigating the process or co-participating in it (e.g. wounds, mind, etc.), whereas the object of cure, a prototypically transitive verb, will be a more serious illness like AIDS or malaria which requires the intervention of others. Similar (lexical-)semantic correlates are pointed out by Keenan (1984). Also Levin & Rappaport (1995:100) mention an observation by A. Kroch on the pair shake (transitive) and shudder (intransitive): "[t]hings that shudder usually can be thought of as having a 'self-controlled body' " whereas things that do not have a "self-controlled" body can only shake. In other words, the notion of independence on behalf of the Medium is reinforced by the collocational patterns of verbs: ergative verbs prototypically occur with an Affected that has the potential of sustaining or self-instigating the process. Levin & Rappaport's use of the notion of selectional restrictions is essentially correct and harmonizes with the present work. However, there is a theoretical incompatibility as to what the criterion is meant to prove. In their derivational approach, Levin & Rappaport use the criterion to support the claim that the causative structure is basic and the non-causative derived via a process of 'detransitivization'. Supporting evidence is the fact that the causative construction imposes less stringent restrictions on its arguments. Cognitive grammar rejects the derivational view on grammar, since it is committed to the idea that "overt structures cannot be derived from hypothetical 'underlying' structures having a substantially different character" (Langacker 1987:54). Unfortunately, Levin & Rappaport's derivational view may lead to descriptive and explanatory inaccuracies. In their view, selectional restrictions on the "subject of the intransitive variant" explain why 'detransitivizations' may be blocked, e.g. *Your mind will open. If this is so, then how is one to explain that slotfillers that do not have the proper selectional restrictions still have the potential to become the subject of such a non-effective construction. For


237

example, in a (non-trivial) SF-series by Julian May which I have read recently, direct communication via the mind was considered possible. The novels abounded in counterexamples to (2), presenting conceptualizations of minds opening, showing or telling. While perhaps more 'creative' than 'ordinary' usage (if this distinction is valid at all), thesec o n s t r u a l sare perfectly sensible, in contrast to what is possible with core transitive verbs. As a rule, the latter do not have this potential, and consequently cannot occur in an ergative noneffective construction. A construction like Tony murdered cannot be understood as an ergativized construction with Tony realizing the role of the Medium. In short, while at the phenotypical level the ergative and transitive effective largely converge, the different semantics of both should be clear. Various elements of language use confirm the cognitive reality of the two models. Further evidence comes from the different constructional extensions which may or may not co-occur with lexical modulations. Both the diachronic changes and idiosyncratic paradigmatic reinterpretations discussed in various places of this book have revealed that speakers operate with the two paradigms. This view once more departs from Levin & Rappaport (1995) when they discuss the possible ergativization of intransitives, as in (3)a. The baby burped. (Smith 1978) The nurse burped the baby. (4)a. The patient bled. (Levin & Rappaport 1995) b. The doctor bled the patient. as "spurious causative pairs", i.e. involving "two distinct verb meanings—one of them causative—that are not derivationally related" (1995:115). However, the descriptions in Chapter 4 have clearly indicated that such constructional changes are not at all spurious, but are the result of the creative exploitation of the two causative models. Language users have a tacit understanding of the semantics of the transitive and ergative models and may 'creatively' stretch the constructional possibilities of a verb in response to the contextual specifications of the usage event. Creativity is here to be understood as comprising the rulegoverned generation of novel structures as well as the elaboration of existing conceptual units (cf. Langacker 1987:56-98). Clearly, this need not be a conscious process nor is it necessarily an individual one. As Langacker puts it, "[l]inguistic creativity is best examined not within the confines of a restricted, self-contained grammar, but rather in the overall context of human knowledge, judgement, and problem-solving ability" (1987:73). For example, it appears that the relatively specific (i.e. verb-bound) constructional shift for abort, a partial transitivization, is motivated within the ideology of the larger speech community (see Chapter 7). Without doubt, such

238


linguistic creativity may at times be highly context-sensitive, i.e. tied to a specific speaker in a given situation confronted with a particular coding problem. In addition the idiosyncratic ergative uses of die mentioned in Chapter 4, section 4.3.1, we can mention an example noted by Halliday (1976:169), an exchange between two children, one holding a small animal: Look it's shivering—Well, you're shivering it. Quite unexpectedly, the second speaker conceptualizes the originally uninstigatable process of shivering as instigated by the first speaker. While likely to be commented upon (in 'proper' English something cannot, as we all know, be 'shiverable') the sentence is clearly understandable. Even Levin & Rappaport mention such an example (with an ergative use of intransitive deteriorate), yet continue to regard it as spurious. This is all the more surprising given that in some cases, patterns of ergativization can become quite productive, as for example the earlier mentioned ergativization of -ize verbs and that of intransitive processes involving machines and technical devices. In other words, such new constructional possibilities may come to constitute well-entrenched units within the grammar. However, entrenchment is neither permanent nor absolute. It is non-absolute given that some units may be more entrenched than others (whence possibly Levin and Rappaport's conjecture that the said ergativizations are spurious). It is non-permanent given that entrenchment may change over time. For example, the ergativization of the transitive SUFFOCATE verbs strangle, throttle and smother, which in the 16th till the 19th century may have been relatively entrenched (although we should be careful here), is no longer sanctioned in present-day English grammar. It can also be noted that while in present-day English the ergativization of many -ize verbs is quite productive, in Jespersen's time it may not yet have been so. On the non-effective use of materialize he says, "now beginning to be common, from US" (1927:334). By way of conclusion, it can be reiterated that the ergative and transitive paradigms do indeed shape the grammar of causative constructions, but that the semantics of the paradigmatically determined construction interacts in nontrivial ways with the semantics of the instantiating verbs. The analyses offered in this book have only scratched the surface of the issue, yet I believe that it clearly has made a case for more nuanced pronouncements on verbal semantics as well as on what is traditionally called the verb's 'transitivity'. Clearly, further research is called for, and the next section offers a short exploration into avenues for such research. 8.3 Further research By combining the paradigmatic perspective with a lexical one, some of the problems raised in the beginning of this book and in the course of the


239

investigation have received a coherent account. At the same time, other issues have not yet been explained or new questions have emerged that require further research. Overall, my view is that much more empirical investigation is called for, given that what is being investigated is not some inherent decontextualized set of characteristics of verbs, but their lexical and constructional flexibility in specific, highly contextualized construals. A wide empirical investigation is indispensable to discover tendencies in language. From Sinclair's experience, "language looks rather different when you look at a lot of it at once" (1991:100). A first extension would be one to a lexically more differentiated corpus. Our analyses have basically been restricted to the field of killing, whose specific lexical properties were shown to influence certain aspects of constructional behaviour. For instance, our analysis of the middle construction and 'middable' verbs would greatly benefit such an extension to other lexical fields to come to more generalized understanding of constraints on a verb's 'middable' character. Another area concerns further investigation into the prototype of the transitive paradigm. For the MURDER verbs, the prototypical structure was shown to incorporate both goal-direction and goal-achievement, thus slightly modifying Davidse's (1991) claim yet confirming Nishimura's (1993). However, the salient incorporation of goal-achievement in the case of the MURDER verbs may be a consequence of their 'lexical' hook-up with intentionality, and might not be so representative for the entire class of transitive verbs, many of which are more general verbs behaving similarly to kill (e.g. make or hit). Wider research could also reveal further constructional differences between the transitive and ergative paradigms that have not been observed so far. For example, verbs of killing indiscriminately have a strong hook-up with a volitional Agent, although perhaps less so for the prototypical ergative verbs. If it is indeed the case that the ergative paradigm has a less stringent commitment to a volitional Instigator, as mentioned in Davidse (1991) and as mildly reflected in some of our corpus material, this could be proven more substantially with an extended corpus. Secondly, the ergative verbs in our field often occur in hyperbolical uses and, related to it, in the progressive aspect, excluding to endpoints of the process. Is this a general tendency of ergative verbs, and if so, how can this be explained? Is there a semantic difference between these hyperbolical uses and those common with transitive kill? While we have sketched some preliminary explanations for both phenomena, further research is clearly necessary. Extensive computerized analysis is bound to be worthwhile for the analysis of collocational patterns. While my differentiation of SUFFOCATE verbs has been based on a manual analysis of collocations, Stubbs (1995) manages to analyse more than 38,000 occurrences of the item cause with the help of special

240


software that can determine collocational patterns. Consequently, he obtains much more reliable and accurate results. I am committed to the idea that collocational analysis is a key to discerning subtle differences in conceptual centres of lexical items. At the same time, it will prove relevant for the distinction between the transitive and ergative constructions, which, as already said, tend to project different 'selection restrictions'. Especially a contrastive analysis of effective constructions of verbs which are lexically near-synonymous yet paradigmatically different may reveal the more general patterns that underlie actual usages. Another phenomenon where extensive contrastive analysis is most relevant is that of the -er derivation, along lines similar to our pilot study in Chapter 5. First, the overall distribution of the transitive and ergative paradigms has to be determined to draw valid conclusions concerning the 'ecological niche' of the -er suffix, i.e. the transitive paradigm. Secondly, if the transitive paradigm is indeed the homeground of the derivation, this should be reflected in a contrastive study of transitive and ergative based derivations. Once again, an investigation of-er formations of near-synonymous verb pairs (or larger sets of verbs) that differ in their prototypically transitive or ergative character, e.g. cure/heal or break/smash, should prove most relevant. If the hypothesis concerning the transitive niche of the derivation is correct, it is to be expected that it is most entrenched for the transitive verb in the pair. At the same time, one has to consider in which context either agentive nominal occurs. For example, when does one use cracker and when smasher? Is there any systematicity in possibly different usages? What type of Agent is each time envisaged? What type of Affected? In other words, do the derivations retain the collocational patterns typical of the AG-PROC-AF constellations? These are but some of the most pressing questions to be answered in a lexical-paradigmatic analysis. Clearly, they are not restricted to the agentive derivation, but pertain to other morphological processes as well, such as the -able derivation or reprefixation. Overall, this book has offered a detailed semasiological study of lexical and constructional variation. That is, starting from specific attestations, it has tried to map out and explain the semantic and constructional range of verbs of killing. The semasiological perspective should be complemented by an onomasiological one, which would look at the data from the other end. That is, it would take as its starting point a particular concept and examine how this concept is expressed (see e.g. Geeraerts et al. 1994 for such an analysis within the framework of Cognitive Grammar). Our limited contrastive analysis of kill versus die in the coding of accidents and diseases is an example of such an onomasiologically oriented study. The data on the SUFFOCATE verbs unequivocally revealed the experiential basis of ergativity and an experientially motivated alignment of the


241

distinction transitive (or effective) versus ergative (or non-effective) with that between external versus internal causes. An onomasiological perspective should confirm these findings. Similarly, an onomasiological study of coding variation may further confirm my observations that the speaker's ideology plays an important role in how an abortion event is coded. Also our diachronic analyses supporting Halliday's suggestion of an ergativization in (Early) Modern English could certainly benefit from such an onomasiological perspective on both diachronic and contemporary data. The most puzzling question in this respect is what motivates this global ergativization process. In our analyses, the ergativization of the SUFFOCATE verbs, of starve, hunger and abort were all shown to be well-motivated. This corroborates the general thesis underlying this book that constructional flexibility should not be regarded in isolation but always be related to the semantics of the verb itself. Nevertheless, the global ergativization which supposedly started somewhere in the 16th century and has supposedly affected a large part of the English vocabulary remains a phenomenon for which we can at present not yet offer a general explanation. Further analysis of extensive data may reveal how productive the ergativization has been and in what context it has occurred. Such an analysis clearly starts from the view that language is not as an isolated phenomenon but an exponent of our experience and interaction with the world. Despite the fact that definitive answers have not always been provided, I believe that this book has offered a fruitful exploration into the complex interplay between verbal and constructional meaning. As is now evident, a verb's constructional potential cannot be stated as fixed and stable but depends on specific usage events. It is these usage events that form the locus of linguistic creativity and language change. I hope that I have been able to lay the methodological foundations, however modest, for further research into this domain.

REFERENCES Allerton, D.J. 1975. "Deletion and Proform Reduction". Journal of Linguistics 11.213-237. Apresjan, Jurij D. 1973. "Regular Polysemy". Journal of Linguistics 142.5-32. Bell, Allan. 1991. The Language of News Media. Oxford: Blackwell. Bolinger, Dwight. 1975. Aspects of Language. New York: Harcourt-Brace Jovanovich. Bresnan, Joan. 1981. The Mental Representation of Grammatical Relations. Cambridge, Mass.: MIT Press. Brook, Stella. 1965. The Language of the Book of Common Prayer. London: Deutsch. Brugman, Claudia M. 1983. The Story of 'over'. Bloomington, Indiana: Indiana University Linguistics.Club. Burzio, Luigi. 1986. Italian Syntax. A Government-Binding Approach. Dordrecht: D. Reidel. Casad, Eugene. 1995. "Seeing It in More than One Way". Language and the Cognitive Construal of the World, ed. by John Taylor & Robert E. MacLaury, 23-49. Berlin & New York: Mouton de Gruyter. Clark, Eve V. and Herbert H. Clark. 1979. "When Nouns Surface as Verbs". Language 55.767-811. Comrie, Bernard. 1973. "The Ergative: Variations on a Theme". Lingua 32.239-253. . 1978. "Ergativity". Syntactic Typology. Studies in the Phenomenology of Language, ed. by W.P. Lehmann, 329-394. Austin: University of Texas Press. . 1981. Language Universals and Linguistic Typology. Oxford: Blackwell. Coulson, Seana. 1992. "Is Incest Best? The Role of Pragmatic Scales and Cultural Models in Abortion Rhetoric". Center for Research in Language Newsletter 7.2. Cowie, A.P. 1982. "Polysemy and the Structure of Lexical Fields". Nottingham Linguistic Circular 11.51-64. Croft, William A. 1991. Syntactic Categories and Grammatical Relations. Chicago: University of Chicago Press.

244 _______


. 1993. "The Role of Domains in the Interpretation of Metaphors and Metonymies". Cognitive Linguistics 4.335-370. _______ . 1994. "The Semantics of Subjecthood". Subjecthood and Subjectivity. ed. by Marina Yaguello, 29-75. Paris & London: Ophrys & Institut français du Royaume-Uni. Cruse, David Α. 1972. "A Note on English Causatives". Linguistic Inquiry 3.522-528. _______ . 1973. "Some Thoughts on Agentivity". Journal of Linguistics 9.281297. _______ . 1982. "On Lexical Ambiguity". Nottingham Linguistic Circular 11.6580. _______ . 1986. Lexical Semantics. Cambridge: Cambridge University Press. Cumps, Christophe. 1990. The Semantics of the English Suffix -tide. A Cognitive Approach. Unpub. M.A. Thesis, K.U.Leuven. Davidse, Kristin. 1987. "M.A.K. Halliday's Functional Grammar and the Prague School". Functionalism in Linguistics. ed. by René Dirven & Vilém Fried, 39-79. Amsterdam & Philadelphia: John Benjamins. _______ . 1991. Categories of Experiential Grammar. Unpub. Ph.D. Dissertation, K.U. Leuven. _______ . 1992. "Transitivity/Ergativity: The Janus-Headed Grammar of Actions and Events". Advances in Systemic Linguistics. ed. by Martin Davies & Louise Ravelli, 105-135. London: Pinter. _______ . 1994. "Language and World View in the Poetry of G.M. Hopkins". Acknowledged Legislators. Essays on English Literature in Honour of Herman Servotte. ed. by Ortwin De Graef, Vic Doyen, Erik Hertog, Guido Latré & Hedwig Schwall, 104-118. Kapellen: Pelckmans. Davidson, Donald. 1980. Essays on Actions and Events. Oxford: Oxford University Press. Deane, Paul. 1987. "English Possessives, Topicality, and the Silverstein Hierarchy". Berkeley Linguistics Society 9.37-46. _______ . 1988. "Polysemy and Cognition". Lingua 75.325-361. _______ & Rebecca S. Wheeler. 1984. "On the Use of Syntactic Evidence in the Analysis of Word Meaning". Papers from the Parasession on Lexical Semantics. Chicago Linguistic Society 20.95-106. Dé Caluwe, Johan. 1995. "Categoriale polysemie en familiegelijkenis. Deverbiaal -er revisited". Tabu 25.3-13. Declerck, Renaat. 1991. A Comprehensive Descriptive Grammar of English. Kaitakusha: Tokyo.

REFERENCES

245

De Coster, S. 1995. "Contrastive Study: clauses with cure versus clauses with heal in the LDC, the WSJ and the NOVEL Corpus". Unpub. term paper, K.U.Leuven. DeLancey, Scott. 1981. "An Interpretation of Split Ergativity and Related Phenomena". Language 57.626-57. _______ . 1982. "Aspect, transitivity, and viewpoint". Tense-Aspect: Between Semantics and Pragmatics, ed. by Paul J. Hopper, 167-183. Amsterdam & Philadelphia: John Benjamins. _______ . 1983. "Agentivity and Causation: Data from Newari". Berkeley Linguistics Society 9.54-63. _______ . 1984a. "Transitivity and Ergativity Case in Lhasa Tibetan". Berkeley Linguistics Society 10.131-140. _______ . 1984b. "Notes on Agentivity and Causation". Studies in Language 8.181-213. _______ . 1985. "Agentivity and Syntax". Papers from the Parasession on Causatives and Agentivity. Chicago Linguistic Society 21.1-12. _______ . 1986. "Evidentially and Volitionality in Tibetan". Evidentiality: The Linguistic Coding of Epistemology. ed. by Wallace Chafe & Johanna Nichols, 203-213. Norwood, NJ: Ablex. _______ . 1987. "Transitivity in Grammar and Cognition". Coherence and Grounding in Discourse, ed. by Russell S. Tomiin, 53-68. Amsterdam & Philadelphia: John Benjamins. _______ . 1990. "Ergativity and the Cognitive Model of Event Structure in Lhasa Tibetan". Cognitive Linguistics 1.289-321. _______ . 1991. "Event Construal and Case Role Assignment". Berkeley Linguistics Society 17.338-353. Dillon, George L. 1974. "Some postulates characterizing volitive NPs". Journal of Linguistics 10.221-233. Dirven, René. 1993. "Dividing up Physical and Mental Space into Conceptual Categories by means of English Prepositions". The Semantics of Prepositions. From Mental Processing to Natural Language Processing. ed. by Cornelia Zelinsky-Zwibbelt, 73-97. Berlin & New York: Mouton de Gruyter. Dixon, Robert M.W. 1979. "Ergativity". Language 55.59-138. _______ . 1994. Ergativity. (= Cambridge Studies in Linguistics 69) Cambridge: Cambridge University Press. Dowty, David R. 1972. "On the Syntax and Semantics of the Atomic Predicate Cause". Chicago Linguistic Society 8.62-74.

246


Emanatian, Michele. 1993. "Figure-Ground Reversal in Grammar". Paper presented at the Third International Cognitive Linguistics Association Conference, Leuven, 18-23 July. Fagan, Sarah M.B. 1988. "The English Middle". Linguistic Inquiry 19.181-203. Faltz, L. 1977. Reflexivization: A Study in Universal Syntax. Unpub. Ph.D. Dissertation, University of California, Berkeley. Fawcett, Robin P. 1980. Cognitive Linguistics and Social Interaction. Towards an Integrated Model of a Systemic Functional Grammar and the Other Components of a Communicating Mind. Exeter & Heidelberg: Exeter University Press & Groos Verlag. Fellbaum, Christiane. 1985. "Adverbs in Agentless Actives and Passives". Papers from the Parasession on Causatives and Agentivity. Chicago Linguistic Society 21.21-31. _______ . 1988. "On the 'Reflexive Middle' in English". Chicago Linguistic Society 25.123-132. _______ . 1990. "English Verbs as a Semantic Net". International Journal of Lexicography 3.278-301. _______ & Judy Kegl. 1989. "Taxonomic Structures and Cross-Category Linking in the Lexicon". Proceedings of the Eastern States Conference on Linguistics, 93-104. _______ & Anne Zribi-Hertz. 1989. "La Construction Moyenne en Français et en Anglais: Etude de Syntaxe et de Sémantique Comparées". Recherches Linguistiques de Vincennes 18.19-57. Feltenius, Leif .1977. "Intransitivizations in Latin". Acta Universitatis Upsaliensis. (= Studia Latina Upsaliensia 9), Stockhom: Almqvist & Wibell International. Feyaerts, Kurt. 1997. Die Bedeutung der Metonymie als konzeptuellen Struktur-prinzips. Eine kognitiv-semantische Analyse deutscher Dummheitsausdrücke. Unpub. PhD. Dissertation, K.U.Leuven. Fiengo, Robert 1980. Surface Structure and The Interface of Autonomous Components. Cambridge, Mass.: Harvard University Press. Fillmore J. Charles. 1968. "The Case for Case". Universals in Linguistic Theory, ed. by Emmon Bach & Robert T. Harms, 1-90. New York: Holt. _______ . 1972. "Subjects, Speakers, and Roles". The Semantics of Natural Language ed. by Donald Davidson & Gilbert Harman, 1-24. Dordrecht: Reidel. _______ . 1977. "The Case for Case Reopened". Grammatical Relations. (= Syntax and Semantics 8) ed. by Peter Cole & Jerrold M. Sadock, 59-81. New York: Academic Press.

REFERENCES _______

247

. 1985. "Frames and The Semantics of Understanding". Quaderni di Semantica 6.222-254. _______ . 1986. "Pragmatically Controlled Zero Anaphora". Berkeley Linguistics Society 12.95-107. _______ . 1988. "The Mechanisms of 'Construction Grammar' ". Berkeley Linguistics Society 14.35-55. _______ , Paul Kay & Mary C. O'Connor. 1988. "Regularity and Idiomaticity in Grammatical Constructions: The Case of Let Alone". Language 64.501-538. Fodor, Jerry A. 1970. "Three Reasons for Not Deriving Kill from Cause to Die". Linguistic Inquiry 1.429-438. Geeraerts, Dirk. 1988. "Where Does Prototypicality Come From?". Topics in Cognitive Linguistics, ed. by Brygida Rudzka-Ostyn, 207-229. Amsterdam & Philadelphia: John Benjamins. _______ . 1993. "Vagueness's Puzzles, Polysemy's Vagaries". Cognitive Linguistics 4.223-272. _______ . 1994. "Generalising Inductive Generalization". Perspectives on English. Studies in Honour of Professor Emma Vorlat. ed. by Keith Carlon, Kristin Davidse & Brygida Rudzka-Ostyn, 128-139. Leuven: Peeters. _______ . & S. Grondelaers. 1991. "Looking Back at Anger: Cultural Traditions and Metaphorical Patterns". K.U.Leuven Preprint 133. _______ , Peter Bakema & Stefan Grondelaers. 1994. The Structure of Lexical Variation: Meaning, Naming, and Context, Berlin & New York: Mouton de Gruyter. Gentner, Dedre and France, I. 1988. "The Verb Mutability Effect: Studies of the Combinatorial Semantics of Nouns and Verbs". Lexical Ambiguity Resolution ed. by Steven L. Small Garrison W. Cottrell, & Michael K. Tannenhaus, 343-382. San Mateo: Kaufmann. Geyskens, Sara. 1997. The Ergativization of Intransitives. A LexicoGrammatical Description. Unpub. M.A. Thesis, K.U.Leuven. Givón, Talmy. 1994. "The Pragmatics of De-transitive Voice: Functional and Typological Aspects of Inversion". Voice and Inversion, ed. by Talmy Givón, 3-44 Amsterdam & Philadelphia: John Benjamins. Goldberg, Adele E. 1992. "The Inherent Semantics of Argument Structure: The Case of the English Ditransitive Construction". Cognitive Linguistics 3.3774. _______ . 1995. Constructions. A Construction Grammar Approach to Argument Structure. Chicago: Chicago University Press. Goossens, Louis. 1990. "Metaphtonymy: The Interaction of Metaphor and Metonymy in Expressions for Linguistic Action". Cognitive Linguistics 1.323-340.

248


Gross, Maurice. 1976. "On the Failure of Generative Grammar". Language 55.859-885. Gruber, Jeffrey S. 1985. "Lexical, Conceptual, and Encyclopedic Meaning". Quaderni di Semantica 6.254-267 Guerssel, M., Kenneth Hale, M. Laughren, Beth Levin, & J. White Eagle. 1985. "A Cross-Linguistic Study of Transitivity Alternations". Papers from the Parasession on Causatives and Agentivity. Chicago Linguistic Society 21.48-63. Haiman, John. 1980. "Dictionaries and Encyclopedias". Lingua 50.329-357. _______ . 1983. "Icononic and Economic Motivation". Language 59.781-819. Hale, Kenneth & Samuel J. Keyser. 1987. "A View from the Middle". (= MIT Lexicon Working Papers 10) University of Massachusetts. Halliday, M.A.K. 1967a. "Notes on Transitivity and Theme in English 1". Journal of Linguistics. 3.37-81. _______ . 1967b. "Notes on Transitivity and Theme in English 2". Journal oj Linguistics 3.199-244. _______ . 1968. "Notes on Transitivity and Theme in English 3". Journal oj Linguistics 4.179-215. _______ . 1976. System and Function in Language. Selected Papers edited by G Kress. Oxford: Oxford University Press. _______ . 1985. An Introduction to Functional Grammar. London: Arnold. Haspelmath, Martin. 1993. "More on the Typology of Inchoative/Causative Verb Alternations". Causatives and Transitivity, ed. by Bernard Comrie & Maria Polinsky, 87-120. Amsterdam & Philadelphia: John Benjamins. Hawkins, Bruce. 1984. The Semantics of Spatial Prepositions. Unpub. Ph.D. Dissertation, University of California, San Diego. Herskovits, Anette. 1982. Space and the Prepositions in English: Regularities and Irregularities in a Complex Domain. Unpub. Ph.D. Dissertation, Stanford University. _______ 1988. "Spatial Expressions and the Plasticity of Meaning". Topics in Cognitive Linguistics, ed. by Brygida Rudzka-Ostyn, 271-297. Amsterdam & Philadelphia: John Benjamins. Hiltbrunner, O. 1981. Bibliographie zur lateinischen Wortforschung. Bern: Francke Verlag. Hoekstra, Teun. 1992. "Aspect and Theta Theory". Thematic Structure. Its Role in Grammar. ed. by LM. Roca, 145-174. Berlin & New York: Foris Publications. _______ & I. Roberts. 1993. "Middle Constructions in Dutch and English". Knowledge and Language. Vol. II: Lexical and Conceptual Structure, ed. by Eric Reuland & Werner Abraham, 183-220. Dordrecht: Kluwer.

REFERENCES

249

Hopper, Paul J. 1985. "Causes and Affects". Papers from the Parasession on Causatives and Agentivity. Chicago Linguistic Society 21.67-88. _______ & Sandra A. Thompson. 1980. "Transitivity in Grammar and Discourse". Language 56.251-299. Horn, Laurence R. 1980. "Affixation and the Unaccusative Hypothesis". Chicago Linguistic Society 16.134-146. Jaeggli, Osvaldo A. & Nina M. Hyams. 1993. "On the Independence and Interdependence of Syntactic and Morphological Properties: English Aspectual come and go". Natural Language and Linguistic Theory 11.313346. Jespersen, Otto. 1927. A Modern English Grammar on Historical Principles. Heidelberg: Carl Winter. Johnson, Mark. 1987. The Body in the Mind: The Bodily Basis of Meaning, Imagination, and Reason. Chicago: Chicago University Press. Kac, Michael B. 1972a. "Action and Result: Two Aspects of Predication in English". Syntax and Semantics 1. ed. by John P. Kimball, 117-124. New York: Academic Press. _______ . 1972b. "Reply to McCawley". Syntax and Semantics 1. ed. by John P. Kimball, 151-156. New York: Academic Press. Katz, Jerrold J. 1970. "Interpretative Semantics vs. Generative Semantics". Foundations of Language 6.220-259. Kay, Paul. 1990. "Even". Linguistics and Philosophy 13.59-112. Keenan, Edward L. 1984. "Semantic Correlates of the Ergative/Absolutive Distinction". Journal of Linguistics 22.97-223. Kemmer, Suzanne. 1993. The Middle Voice (= Typological Studies in Language 23) Amsterdam & Philadelphia: John Benjamins. _______ & A. Verhagen. 1994. "The Grammar of Causatives and the Conceptual Structure of Events". Cognitive Linguistics 5.115-156. Keyser, Samuel J. & Thomas Roeper. 1984. "On the Middle and Ergative Constructions in English". Linguistic Inquiry 15.381-416. King, Robert T. 1988. "Spatial Metaphor in German Causative Constructions". Topics in Cognitive Linguistics, ed. by Brygida Rudzka-Ostyn, 556-585. Amsterdam & Philadelphia: John Benjamins. Lakoff, George. 1970. Irregularity in Syntax. New York: Holt. _______ . 1977. "Linguistic Gestalts". Chicago Linguistic Society 13.236-287. _______ . 1986. "FrameSemantic Control of the Coordinate Structure Constraint". Chicago Linguistic Society 22.152-167. _______ . 1987. Women, Fire, and Dangerous Things. What Categories Reveal About the Mind. Chicago: Chicago University Press.

250 _______


& Mark Johnson. 1980. Metaphors We Live By. Chicago: Chicago University Press. _______ & Zoltan Kövecses. 1987. "The Cognitive Model of Anger Inherent in Amercian English". Cultural Models in Language and Thought, ed. by Dorothy Holland & Naomi Quinn, 195-221. Cambridge: Cambridge University Press. _______ & John R. Ross. 1972. "A Note on Anaphoric Islands and Causatives". Linguistic Inquiry 3.121-125. Langacker, Ronald W. 1978. "The Form and Meaning of the English Auxiliary". Language 54.853-882. _______ . 1979. "Grammar as Image". Linguistic Notes from La Jolla 6.87-126. _______ . 1982a. "Space Grammar, Analysability, and the English Passive". Language 58.22-80. _______ . 1982b. "Remarks on English Aspect". Tense-Aspect: Between Semantics and Pragmatics, ed. by Paul J. Hopper, 265-304. Amsterdam & Philadelphia: John Benjamins. _______ . 1984. "Active Zones". Berkeley Linguistics Society 10.172-188. _______ . 1985. "Observations and Speculations on Subjectivity". Iconicity in Syntax, ed. by John Haiman, 108-150. Amsterdam & Philadelphia: John Benjamins. _______ . 1986. "Abstract Motion". Berkeley Linguistics Society 12.455-471. _______ . 1987a. Foundations of Cognitive Grammar. Vol. I: Theoretical Prerequisites. Stanford: Stanford University Press. _______ . 1987b. "Grammatical Ramifications of the Setting/Participant Distinction". Berkeley Linguistics Society 13.383-394. _______ . 1987c. "Nouns and Verbs". Language 63.53-94. _______ . 1988a. "An Overview of Cognitive Grammar". Topics in Cognitive Linguistics, ed. by Brygida Rudzka-Ostyn, 3-48. Amsterdam & Philadelphia: John Benjamins. _______ . 1988b. "A View of Linguistic Semantics". Topics in Cognitive Linguistics, ed. by Brygida Rudzka-Ostyn, 49-90. Amsterdam & Philadelphia: John Benjamins. _______ . 1988c. "The Nature of Grammatical Valence". Topics in Cognitive Linguistics, ed. by Brygida Rudzka-Ostyn, 91-125. Amsterdam & Philadelphia: John Benjamins. _______ . 1988d. "A Usage-Based Model". Topics in Cognitive Linguistics, ed. by Brygida Rudzka-Ostyn, 127-161. Amsterdam & Philadelphia: John Benjamins. _______ . 1988e. "Case and Grammatical Relations in Cognitive Grammar (with special reference to Newari)". Linguistic Notes from La Jolla 14.57-94.

REFERENCES _______

251

. 1990a. "Subjectification". Cognitive Linguistics 1.5-39. . 1990b. "Cognitive Grammar: the Symbolic Alternative". Studies in the Linguistic Sciences 20.3-30. _______ . 1991a. Concept, Image, and Symbol. Berlin & New York: Mouton de Gruyter. _______ . 1991b. Foundations of Cognitive Grammar. Vol II: Descriptive Application. Stanford: Stanford University Press. _______ . 1992. "The Symbolic Nature of Cognitive Grammar: The Meaning of of and of-periphrasis". Thirty Years of Linguistic Evolution, ed. by Martin Pütz, 483-502. Amsterdam & Philadelphia: John Benjamins. _______ . 1993. "Reference-Point Constructions". Cognitive Linguistics 4.1-38. _______ . 1995. "Raising and Transparancy". Language71.1-62. Lehrer, Adrienne. 1974. "Homonymy and Polysemy: Measuring Similarity of Meaning". Language Sciences 32.33-39. Lemmens, Maarten. 1993. "The Conceptual Impact of Killing". Paper presented at the Third International Cognitive Linguistics Association Conference, Leuven, 18-23 July. _______ . 1994. "The Transitive-Ergative Interplay in a Polysemic Complex: The Issue of Abortion". Perspectives on English. Studies in Honour of Professor Emma Vorlat. ed. by Keith Carlon, Kristin Davidse & Brygida RudzkaOstyn, 345-385. Leuven: Peeters. _______ . 1997. "The Influence of World Conception on Transitivity and Ergativity: a Case Study". Lexical and Syntactic Constructions and the Construction of Meaning, ed. by Eve Sweetser, Kee Dong Lee & Marjolijn Verspoor, 363-382. Amsterdam & Philadelphia: John Benjamins. Levin, Beth. 1993. English Verb Classes and Alternations. A Preliminary Investigation. Chicago: Chicago University Press. _______ . & Malka Rappaport Hovav. 1988. "Nonevent -er Nominals". Journal of Linguistics 26.1067-1083. _______ . 1991. "Wiping the Slate Clean: A Lexical Semantic Exploration". Cognition 41.123-151. _______ . 1992. "The Lexical Semantics of Verbs of Motion: The Perspective from Unaccusativity". Thematic Structure. Its Role in Grammar, ed. by I.M. Roca, 247-269. Berlin & New York: Foris Publications. _______ . 1995. Unaccusativity. At the Syntax-Lexical Semantics Interface. (= Linguistic Inquiry Monographs 26) Cambridge, Mass.: MIT Press. Lindner, Susan. 1982. "What Goes Up Doesn't Necessarily Come Down: The Ins and Outs of Opposites". Chicago Linguistic Society 18.305-323. Lipka, Leonhard. 1986. "Homonymie, Polysemie oder Ableitung im heutigen Englisch". Zeitschrift für Anglistik und Amerikanistik 34.128-138.

_______

252


Löschhorn, Β. 1976. "Die Grundbedeutung von orior". Museum Helveticum 33.105-112. Lyons, John. 1977. Semantics. Vol. I & II. Cambridge: Cambridge University Press. Marantz, Alec P. 1984. On the Nature of Grammatical Relations. Cambridge, Mass.: MIT Press. McCawley, James 1968. "Lexical Insertion in a Transformational Grammar without Deep Structure". Chicago Linguistic Society 4.71-80. Mignot, Xavier. 1969. Les verbes dénominatifs latin (= Etudes et Commentaires 71) Paris: Librarie C. Klincksieck. Nishimura, Yoshiki. 1993. "Agentivity in Cognitive Grammar". Conceptualizations and Mental Processing in Language, ed. by Richard A. Geiger & Brygida Rudzka-Ostyn, 488-530. Berlin & New York: Mouton de Gruyter. Mittwoch, Anita. 1971. "Idioms and Unspecified NP Deletion". Linguistic Inquiry 2.255-259. _______ . 1982. "On the Difference between Eating and Eating Something: Activities versus Accomplishments". Linguistic Inquiry 13.113-122. Nunberg, Geoffrey. 1978. The Pragmatics of Reference, Bloomington, Indiana: Indiana Linguistic Club. _______ . 1979. "The Non-Uniqueness of Semantic Solutions: Polysemy. Linguistics and Philosophy 2.143-184. Perlmutter, David M. 1978. "Impersonal Passives and the Unaccusative Hypothesis". Berkeley Linguistics Society 4.157-89. Pinker, Stephen. 1989. Learnability and Cognition: The Acquisition of Argument Structure. Cambridge, Mass.: MIT Press. Plank, Frans., ed. 1979. Ergativity. Towards a Theory of Grammatical Relations. New York: Academic Press. Quirk, Randolph, Sidney Greenbaum, Geoffrey Leech & Jan Svartvik. 1985. A Comprehensive Grammar of the English Language. London: Longman. Rapoport, T.R. 1993. "Verbs in depictives and resultatives". Semantics and the Lexicon, ed. by James Pustejovsky, 163-184. Dordrecht: Kluwer. Rice, Sally 1987a. Towards a Cognitive Model of Transitivity, Unpub. Ph.D. Dissertation, University of California, San Diego. _______ . 1987b. "Towards a Transitive Prototype: Evidence from Some Atypical English Passives". Berkeley Linguistics Society 13.422-434. _______ . 1988. "Unlikely Lexical Entries". Berkeley Linguistics Society 14.202212.

REFERENCES

253

Rudzka-Ostyn, Brygida. 1988. "Semantic Extension into Verbal Communication". Topics in Cognitive Linguistics, ed. by Brygida RudzkaOstyn, 507-553. Amsterdam & Philadelphia: John Benjamins. _______ . 1994. "Metaphor, Schema, Invariance: the Case of Verbs of Answering". Perspectives on English. Studies in Honour of Professor Emma Vorlat. ed. by Keith Carlon, Kristin Davidse & Brygida RudzkaOstyn, 408-447. Leuven: Peeters. Ryder, Mary E. 1991a. "From writere to in-the-parker: an analysis of -er nominals in English". Lecture presented at the Second International Cognitive Linguistics Association Conference, Santa Cruz, California, 29 July - 2 August. _______ . 1991b. "Mixers, Mufflers and Mousers: The Extending of the -Er Suffix as a Case of Prototype Reanalysis". Berkeley Linguistics Society 17.299-311. _______ . 1997. "Development of the English -er suffix: a corpus study". Lecture presented at the Fifth International Cognitive Linguistics Association Conference, Amsterdam, 14-19 July. Schlesinger, I.M. 1989. "Instruments as Agents: On the nature of semantic relations". Journal of Linguistics 25.189-210. Shibatani, Masayoshi. 1972. "Three Reasons for Not Deriving 'Kill' from 'Cause to Die' in Japanese". Syntax and Semantics 1. ed. by John P. Kimball, 125-137. New York: Academic Press. _______ . 1973. "Lexical versus Analytical Causatives in Korean". Journal of Linguistics 9.281-297. _______ . 1976. "The Grammar of Causative Constructions: A Conspectus". The Grammar of Causative Constructions (=Syntax and Semantics 6) ed. by Masayoshi Shibatani, 1-40. New York: Academic Press. Shopen, Timothy. 1977. "Caught in the Act". Chicago Linguistic Society 7.254-263. Sinclair, John M. 1991. Corpus, Concordance, Collocation. Oxford: Oxford University Press. Smith, Carlota S. 1978. "Jespersen's 'Move and Change' Class and Causative Verbs in English". Linguistic and Literary Studies in Honor of Archibald A. Hill. Vol. II: Descriptive Studies, ed. by M.A. Jazayery, E.C. Polome, and W. Winter, 101-109. The Hague: Mouton. Starosta, Stanley. 1978. "The One per Cent solution". Valence, Semantic Case and Grammatical Relations, ed. by Werner Abraham, 459-576. Amsterdam & Philadelphia: John Benjamins. Stubbs, Michael. 1995. "Collocation and Semantic Profiles". Functions of Language 2.23-55.

254


Talmy, Leonard. 1985a. "Force-Dynamics in Language and Thought". Papers from the Parasession on Causatives and Agentivity. Chicago Linguistic Society 21.293-337. _______ . 1985b. "Legalization Patterns: Semantic Structure in Lexical Forms". Language Typology and Syntactic Description. Vol. III: Grammatical Categories and the Lexicon, ed. by Timothy Shopen, 57-149. Cambridge: Cambridge University Press. _______ . 1991. "Path to Realization: a Typology of Event Conflation". Berkeley Linguistics Society 17.480-519. Taylor, JohnR. 1989. Linguistic Categorization. Oxford: Clarendon. Tuggy, David. 1988. "Náhuatl Causative/Applicatives in Cognitive Grammar". Topics in Cognitive Linguistics, ed. by Brygida Rudzka-Ostyn, 587-618. Amsterdam & Philadelphia: John Benjamins. _______ . 1993. "Ambiguity, Polysemy, and Vagueness". Cognitive Linguistics 4.273-290. Vandeloise, Claude. 1991. Spatial Prepositions: A Case Study from French. Chicago: Chicago University Press. Van Oosten, Jeanne. 1977. "Subjects and Agenthood in English". Chicago Linguistic Society 13.459-71. Vendler, Zeno. 1984. "Adverbs of Action". Papers from the Parasession on Lexical Semantics. Chicago Linguistic Society 20.297-307. Whorf, Benjamin Lee. 1956. Language, Thought and Reality. Selected Writings of Benjamin Lee Whorf. ed. by J.B. Carroll. Cambridge, Mass.: MIT Press. Wierzbicka, Anna. 1980. Lingua Mentalis: the Semantics of Natural Language. New York: Academic Press. _______ . 1988. The Semantics of Grammar. (= Studies in Language Companion Series 8) Amsterdam & Philadelphia: John Benjamins. _______ . 1991. Cross-Cultural Pragmatics: The Semantics of Human Interaction, Berlin & New York: Mouton de Gruyter . _______ . 1992. Semantics, Culture and Cognition: Universal Human Concepts in Culture-Specific Configurations. Oxford: Oxford University Press. Williams, Edwin. 1981. "On the Notions 'Lexically Related' and 'Head of a Word' ". Linguistic Inquiry 12.245-74. Wojcik, Richard H. 1976. "Where Do Instrumental NPs Come From?". The Grammar of Causative Constructions (= Syntax and Semantics 6) ed. by Masayoshi Shibatani, 166-180. New York: Academic Press. Zwicky, Arnold M. & Jerrold M. Sadock. 1975. "Ambiguity Tests and How to Fail Them". Syntax and Semantics 4. ed. by John P. Kimball, 1-36. New York: Academic Press.

DICTIONARIES Collins Cobuild English Language Dictionary 1987. London: HarperCollins. Collins Cobuild Dictionary of Phrasal Verbs 1989. London/Glasgow: Collins. Kurath, H. & S.M. Kuhn, eds. 1956. Middle English Dictionary. Ann Arbor, Mich.: University of Michigan Press. Latham, R.E., ed. 1975. Dictionary of Medieval Latin from British Sources. Oxford: Oxford University Press. The New Lexicon Webster's Dictionary 1978. New York: Lexicon Publications, Inc. The Oxford English Dictionary. 1973. Oxford: Clarendon Press. The Oxford English Dictionary on CD-ROM. 1993. Oxford: Clarendon Press. Partridge, E. 1979. Origins. A Short Etymological Dictionary of Modern English. London: Routledge. Shipley, J.T. 1945. Dictionary of Word Origins. New York: The Philosophical Library. A Supplement to the Oxford English Dictionary, 1972. Oxford: Clarendon Press. Thesaurus Linguae Latinae, Volumen I, 1900. s.l. Webster's 1973. Third New International Dictionary, Springfield, Mass: Merriam Company. Webster's 1978. New Dictionary of Synonyms. Springfield, Mass: Merriam Company.

CORPUS REFERENCES CASS: Contemporary American Short Stories (see section 1.4.2) LDC: Leuven Drama Corpus (see section 1.4.2) NOV: collection of 19th century novels (see section 1.4.2) WSJ: Wall Street Journal (from 25/07 to 02/11/1989) (see section 1.4.2) HHGG: Douglas Adams, Hitchhiker's Guide to the Galaxy, London: Pan Books, 1993. LR: Jack Butler, Living in Little Rock with Miss Little Rock, London: Abacus, 1994. ΑΒΟΚΉΟΝ: postings to the newsgroups alt. abortion and alt.talk,abortion, 1996. sw: Silent Witness, broadcast 28/11/1997 OED: Oxford English Dictionary (reference mostly followed by date) OED on CD-ROM: Oxford English Dictionary on CD-ROM (idem) MED: Middle English Dictionary (reference mostly followed by date) TLL: Thesaurus Linguae Latinae DMLBS: Dictionary of Medieval Latin from British Sources (1975) COLLINS: Collins Cobuild (1987) LDPV: Longman's Dictionary of Phrasal Verbs (1983) WEB: Webster's (1973)

GLOSSARY The present overview identifies the abbreviations and terminology used in this book. For the sake of concenience, they have been grouped in different subheadings. For exact references to where these concepts are explained the reader is referred to the index. 1. Participants transitive Agent (Ac) Actor ergative Agent (Is) Instigator schematizes over Ac and Is (AG) Agent transitive Affected (Go) Goal ergative Affected (ME) Medium schematizes over Go and ME (AF) Affected ActorMedium schematizes overACintrand ME (ACME) transitive pseudo-Affected (RA) Range ergative pseudo-Agent (SE) Setting 2. Construction types effective: two-participant construction; AG-PROC-AF transitive: Ac-PROC-Go, e.g. John killed Mary ergative: IS-PR0C-ME, e.g. Jill drowned Jean non-effective: one-participant construction; ACME-PROC transitive: Ac-PROC e.g. John died ergative: ME-PROC, e.g. Jean drowned pseudo-effective: one-participant construction with pseudo-participant transitive: Ranged construction, AC-PROC-RA, e.g. He died a slow death ergative: Setting construction, SE-PROC-ME, e.g. The house blew afuse intransitive: transitive one-participant construction, Ac-PROC, e.g. John died objectless transitive: transitive effective with omitted object, Ac-PROC-(Go) e.g. Soldiers kill middle: medio-passive construction; AF-PROC-(AG) transitive: Go-PROC-(Ac), e.g. Stale bread cuts easily

258

LEXICAL

ergative: difficulty

PERSPECTIVES ON TRANSITIVITY AND ERGATIVITY ME-PROC-(IS),

e.g. The window opened only with great

3. Miscellaneous construction: the grammatical frames in which a verb may occur, e.g. ergative non-effective, transitive middle, etc. ergativization: process via which a transitive or intransitive verb comes to be used in ergative constructions. For the former this means that the Affected is conceived as possessing the possibility of self-instigating the process; for the latter, that the process can be conceptualized as instigatable. inchoative: term sometimes used in the linguistic literaure to refer to a construction like The glass broke; I will use the term (ergative) noneffective. middable: transitive verbs that allow both the middle construction and the -able derivation. non-causative: traditional term to denote a construction like The glass broke; the two-participant construction is referred to as a causative construction. I will not continue this term but use the term ergative non-effective instead. transittvization: process via which prototypically ergative verbs occur in constructions which no longer instantiate the ergative paradigm as they do not allow the non-effective alternation.

INDEX OF NAMES

A Allerton, D.J., 140 Apresjan, Jurij D., 54

Β Bell, Allan, 18 Bolinger, Dwight, 141 Bresnan, Joan, 187 Brook, Stella, 155 Brugman, Claudia, 16 Burzio, Luigi, 16; 45

Deane, Paul, 29; 54; 63; 101; 103 Declerck, Renaat, 121 DeLancey, Scott, 4; 21; 59; 99; 101-5; 123-26; 172 Dillon, George L., 107 Dirven, René, 123 Dixon, Robert M.W., 4; 28-30; 31; 42 Dowty, David R., 21; 22

E Emanatian, Michèle, 12; 123 Evenepoel, Willy, 197

c

F

Casad, Eugene, 10 Clark, Eve, 57; 114 Clark, Herbert, 57; 114 Clarysse, Willy, 197 Comrie, Bernard, 4; 28 Coulson, Seana, 54; 212-13; 218 Cowie, A.P., 63 Croft, William Α., 10; 20; 44; 54; 79 Cruse, David Α., 63; 86; 99; 109 Cumps, Christophe, 129

Fagan, Sarah B.M., 72; 74; 77 Faltz, L., 73 Fawcett, Robin P., 73 Fellbaum, Christiane, 36; 43; 45; 49; 71-78; 72; 86; 140 Feltenius, Leif, 45; 60 Feyaerts, Kurt, 55 Fiengo, Robert, 72

D Davidse, Kristin, vii; 3; 4; 6; 8; 13; 16; 22; 26; 27; 28; 30; 31; 37; 39-46; 47; 59; 60; 61; 62; 65; 72; 73; 77; 78; 82; 85; 86; 90; 91; 92; 106; 109; 115; 116; 118; 124; 125; 126; 141; 142; 145; 163; 209; 221; 222; 234; 235;239 Davidson, David, 120 De Caluwe, Johan, 99 De Coster, S., 236

Fillmore, Charles J., 9; 40; 100; 140; 141 Fodor, Jerry Α., 21; 22; 23-28 France, Ilene, 49

G Geeraerts, Dirk, vii; 60; 63; 160; 168; 174; 183;207;208; 216; 240 Gentner, Dedre, 49 Geyskens, Sara, 86; 92 Givón, Talmy, 124 Goldberg, Adele E., 5; 9; 16; 38; 232 Goossens, Louis, 54; 195 Grondelaers, Stef, 183 Gross, Maurice, 15

260


Gruber, Jeffrey, 10 Guerssel, M., 40

H Haiman, John, 10; 26 Hale, Kenneth, 36 Halliday, M.A.K., 8; 13-14; 15; 28; 31; 39; 40; 41; 73; 91; 93; 94; 224; 238; 241 Haspelmath, Martin, 40; 44; 59 Hawkins, Bruce W., 16 Herskovits, Annette H., 169; 170 Hiltbrunner, Otto, 199 Hjelmslev, L., 13 Hoekstra, Teun, 21; 72 Hopper, Paul, 39; 43; 124; 186 Horn, Lawrence, 16; 79; 82; 84; 85 Hyams, Nina M., 121

L Lakoff, George, 9; 10; 16; 21; 22; 25; 27; 54; 73; 74; 121; 150; 179; 181; 183 Langacker, Ronald W., 6; 8-13; 10; 11; 13; 14; 15; 20; 27; 28; 30-36; 37; 39; 45; 50; 72; 73; 74; 77; 90; 103; 104; 117; 123; 149-52; 200;237 Latham, RE., 199; 201 Lehrer, Adrienne, 63 Lemmens, Maarten, 191; 203; 207 Levin, Beth, 4; 8; 16; 29; 30; 36-39; 40; 44; 45; 49; 55-58; 60; 66; 72; 79; 82; 86; 90; 92; 128-31; 137; 165; 182; 221; 232; 236; 237; 238 Lindner, Susan, 50; 181; 184 Lipka, Leonhart, 63 Löschhorn, B., 198 Lyons, John, 10; 63; 124

J Jaeggli, Osvaldo Α., 121 Jespersen, Otto, 45; 60; 72; 86; 96; 140; 142; 146; 238 Johnson, Mark, 16; 25; 54; 179; 181

Κ Kac, Michael Β., 117 Katz, Jerrold J.,21;22;25 Kay, Paul, 9 Keenan, Edward L., 236 Kegl, Judy, 45; 140 Kemmer, Suzanne, 72; 73; 119 Keyser, Samuel J., 8; 16; 36; 39; 43; 72; 73; 77; 78; 79; 81; 82; 84; 92; 135; 136; 187; 229 King, Robert T., 25 Kövecses, Zoltan, 183 Kroch, Α., 236

M Marantz, Alec P., 16; 44; 101; 128 McCawley, James, 21; 22; 25; 28 Mignot, Xavier, 199 Mittwoch, Anita, 140

Ν Nishimura, Yoshiki, 99; 101; 104; 105; 116; 124;125;145; 239 Nunberg, Geoffrey, 54; 63

Ρ Partridge, Eric, 114; 153; 197 Perlmutter, David M., 16; 45 Pinker, Steven, 86 Plank, Frans, 28

Q Quirk, Randolph, 121

INDEX OF NAMES

R

Stubbs, Michael, 17; 239

Rapoport, T.R., 79; 80; 228 Rappaport, Malka Hovav, 8; 16; 38; 45; 66; 92; 128-31; 137; 182; 236; 237; 238 Rice, Sally, 35; 38; 39; 106; 140-46; 210; 222; 227;229; 232 Roberts, I., 72 Roeper, Thomas, 8; 16; 39; 43; 72; 73; 77; 78; 79; 81; 82; 84; 92; 135; 136; 187; 229 Ross, John Robert, 27 Rudzka-Ostyn, Brygida, vii; 45; 53; 54; 178 Ryder, Mary E., 73; 103; 104; 105; 120; 128; 137; 139; 229

Τ

S Sadock, Jerrold M., 63 Schlesinger, I.M., 99; 101; 102; 104 Shibatani, Masayoshi, 22; 25; 27 Shipley, J.T., 153; 154; 197; 205 Shopen, Timothy, 121 Sinclair, John, 17; 239 Smith, Carlotta, 25; 40; 73; 75; 79; 86; 92; 228; 237 Starosta, Stanley, 100

261

Talmy, Leonard, 24; 31; 33; 60; 124; 125; 179; 182; 184; 186 Taylor, John R., 10 Thompson, Sandra, 39; 43; 124; 186 Tuggy, David, 51; 63

ν Van Oosten, Jeanne, 73 Vandeloise, Claude, 181 Vendler, Zeno, 25; 43; 73; 76; 235 Verhagen, Arie, 119

W Wheeler, Rebecca S., 29; 101 Whorf, Benjamin, L., 29 Wierzbicka, Anna, 12; 13; 22; 26; 27; 99 Wojcik, Richard H., 101; 128

Ζ Zribi-Hertz, Anne, 71-78; 86 Zwicky, Arnold M., 63

INDEX OF SUBJECTS assassinate, 108-9

A abort, 96-97; 191-219 ergativization, 200-205 etymology, 197-200 metonymical and metaphorical variation, 192-95 onomasiological variation, 214 round-up, 224 transitivization, 208-15 abortivate, 202 absolutive, 28-33; 45 accidental action, 118-19 accommodation, 50; 55; 65 accusative, 28-33; 46; 99; 196; 229 acquire, 79 action chain, 31-32; 33; 72; 103; 211; 219 action verbs, 57-58 active zone, 149-52; 166 Actor, 39-46; 106-27 ActorMedium, 62 Affected, 39-46 Agent, 7; 32; 39-46; 99-106 inanimate Agents, 119-26 versus Cause, 123-26 agentive nominals see -er. agentivity, 7; 99-106 all-nighter, 130 alterable, 84 alternation method, 14; 16; 30; 36-39 ambitransitive verbs, 30 analytical causative, 5; 22-28; 43; 103; 119 anneal, 79 asphyxiate, 159-60 assassinatable, 82

Β behead, 58 billiard-ball model, 31-32 bite, 150 bleed, 86; 92; 114; 234 blush, 92 boil, 43; 232 bounce, 72 break, 23; 30; 33; 36; 40; 133; 232; 236 breaker, 132 bribe, 11 build, 126 builder, 132 burden (verb), 171 burn, 34 burp, 92; 234 burst, 30; 34 butcher, 113-15

c canonical event model, 31-32 case, 28-33 CASS corpus, 18-20 Cause, 22; 123-26; 236 cause to die, 22-28 chair (verb), 59 change, 37; 47; 65 change of state, 36-37; 78-82 changeable, 84 chase, 150 chip, 37 choke, 168-73 choke an engine, 173 choke back, 179-80

INDEX OF SUBJECTS choke down, 178-79 choke off, 181-83 choke on, 168-71 choke out, 180-81 choke up, 183 choke with, 171-73 chômeur, 15 -cide suffix, 129 circumstantial passive, 103; 121-23 clean (verb), 143 cleaner, 130 climb, 90 cognate object, 44 cognitive domain, 9;10 Cognitive Grammar, 8-13 collective noun, 108; 113 collocation, 17; 175; 176; 177; 218; 223; 233; 236;237 compound, 106; 132; 133; 134; 158 conceptual base, 64; 154 Construction Grammar, 9 constructional polysemy, 38; 64-66 cook (verb), 143 corpus description, 18-20 covert grammatical category, 4 crack, 36 cracker, 236 crave, 90 crucify (verb), 57 cryptotype, 29; 46; 231 cure, 233; 236 cut, 34; 107

D decapitate, 58 DECAPITATE verbs, 58

decimate, 128 decollate, 58 dependable, 84 desire (verb), 57

263

deteriorate, 235 detransitivization, 233 die, 92 ergativization of, 91-92 versus kill, 123-26 dipper, 137 do ro-construction, 23 do-test, 73 doublet, 154-55 drink, 38; 142 drinkable, 84 drive, 77; 78; 81; 85; 94; 230 drop, 118 drown, 161-64 drown out, 184-86 drownable, 83; 85 drowner, 135 dry (verb), 143 dummy, 15

E eat, 142 effective construction, 9; 43-45 egocentric view, 103; 212 eliminate, 128 encyclopedic knowledge, 10; 79 entertain, 143 entrenchment, 63; 90; 91; 92; 94; 96; 145; 167; 168; 188; 202; 235; 236 episode, 104; 122; 129; 211 episodic noun, 11; 117 -er (agentive suffix), 43; 127-40; 226-27 Goal-profiling-er, 137-39 Medium-profiling-er, 135-37 erect, 75 ergativization, 20; 62; 67; 77; 86-92-94-96; 97; 127; 129; 136; 139; 153; 156; 157; 167; 173; 188; 192; 196; 200-205; 211; 218; 219; 223; 224; 225; 230; 234; 235; 237 executable, 83

264


execute, 109-11 experience, 32; 47; 59; 67; 94; 97; 151; 152; 157; 220; 237 explode, 101 exterminate, 128

hit, 34; 145 homonymy, 62-66; 168; 218; 225 hunger, 89-91; 225 hyperbole, 55; 90; 116; 152; 164; 223; 236 hyperonymy, 115; 191

F

I

faint, 92 fall, 34; 126 famish, 149 fasten, 11 federalize, 92 figure/ground alignment, 10-12; 198; 200; 208 fill, 171 float, 136 flog, 43 fold, 75 freeze, 34 frequentative -le, 154 fungicide, 129

iconicity, 26 ideology, 67; 110; 111; 234; 237 imagery, 10; 175; 200 inceptive verbs, 44 inchoative, 34. See ergative non-effective inergative, 45 infest, 171 -ing nominalization, 117; 142 instantiation, 12; 49-55; 60-62; 109; 221 instigatability, 86; 92; 214; 224; 225; 230; 231 Instigator, 39-46

G garrotte (verb), 57 Generative Grammar, 14-16 girdle 154 Goal, 39-46 goal-achievement, 116-18 grammatical metaphor, 13 grammaticalization, 130 grinder, 130 grow, 44 gun (verb), 19; 57

H hammer, 142 handle, 154 hang, 96 harden, 26 heal, 233; 236 herbicide, 129

INSTRUMENT verbs, 57

intentionality see volition. interrupt, 142 intonation, 77; 108 intransitive construction, 43-45 -ion nominalization, 117 -ize suffix, 92; 235

J journalese, 18 jump, 85; 86; 126

K kick, 79; 80 kickable, 82 kidnap, 142 kill, 50-55; 115-23 versus die, 123-26 killable, 83 killer, 129; 132 middle-based, 137

INDEX OF SUBJECTS knife (verb), 19; 57 know, 79

L landmark, 12; 123; 149; 150; 181; 182; 184; 185;193 laughable, 84 LDC corpus, 18-20 -le suffix, 154 learn, 79 lexeme, 63; 231 lexical category, 49-55 lexical fields, 55-58 lexicogrammar, 13 like (verb), 39 load (verb), 171 locative, 124; 181; 184 long (verb), 57; 90 look, 39 lose one's life, 19 lynch, 111 lynchable, 82

M machine-gun (verb), 19; 57 maker, 132 march, 86 marked passive, 90-91 markedness, 90; 127; 216 mash, 37 massacre, 113 materialize, 235 Medium, 39-46 melt, 27; 34 metaphor, 5; 16; 49; 51-52; 54; 55; 95; 175; 180;183; 185; 192-95 metaphtonymy, 54; 188; 195 metonymy, 5; 49; 52; 54; 55; 64; 95; 103; 108; 130; 149; 158; 166; 175; 178; 180; 186; 192-95; 198; 202; 204-5; 211

265

middable verbs, 71; 82-85; 146; 226; 227 middle construction, 6; 17; 34; 35; 37; 38; 47; 62; 71-85; 97; 106; 127; 143; 187; 225-26; 227;232 middle-based agentive nominals, 137-39 miscarry, 205-8 motion verbs, 86; 92 muffle, 184 murder, 107-8 MURDER verbs

delineation, 55-56 description, 107-27 overview, 106-7 round-up, 220-22 murderable, 82; 83 murderee, 83 murderous, 108

N no-brainer, 130 nominalization, 11-12; 39; 117; 131; 142; 198 nominative, 28-33; 46; 99; 196; 229 non-effective construction, 9; 43-45 non-instigatability, 86; 88 NOV corpus, 18-20

o objectless ergative construction, 202; 203; 218 objectless transitive construction, 3; 35; 38; 41; 42; 45; 64; 67; 74; 81; 96; 106; 127; 14046; 144; 161; 187; 196; 210; 211; 227 OED corpus, 18-20 off, 181-83 omitted object see objectless transitive construction. on, 168-71 onomasiological variation, 207; 237 open, 34; 37; 47; 65; 133; 232 out, 184-86 outbounce, 187

266


outkill, 187 out-prefix, 187 outrun, 187 overt grammatical category, 4. See also phenotype overwhelm, 184 oxidize, 92

Ρ Part/Whole metonymy, 52; 54; 193; 195 particle, 178-87; 227-28 passivization, 90-91 peel, 34 perish, 19 perishable, 85 phenotype, 29; 231; 234 phonology, 13 phrasal verbs, 178-87; 227-28 pine (away), 146 please, 39 poison (verb), 19; 56; 57; 159 pollute, 171 polygenesis, 216 polysemy, 39; 49-55; 60; 62-66; 96 premeditation, 107; 117; 222 processual predication, 33 productivity, 130 profile/base distinction, 10 progressive aspect, 116; 117; 143; 224; 236 prototype category, 12; 42; 49-55; 60-62; 6364 prowl, 90 pseudo-effective construction, 43-45; 90-91 pseudo-Goal, 44; 91 pseudo-Instigator, 44 pulverize, 37 punch, 107

Q quench, 155;157

R Ranged construction, 44 rape (verb), 142 reaccelerate, 135 read, 73 readable, 84 readjourn, 135 recategorization, 104; 124 reclose, 79 recover, 146 reference point. See active zone register, 18; 111; 160 reharden, 135 Relational Grammar, 14-16 relational predication, 11-12 relock, 79 reopen, 79 reorganize, 135 re-prefixation, 43; 79; 237 representativity, 17; 105 resemble, 39 reshut, 79 role archetypes, 31-32 rope (verb), 57 row, 90 run, 85; 86; 126; 209

S save, 41 saw (verb), 142 schema, 12 see also schematization. schematicity, 146; 227 schematization, 12; 15; 31; 41; 49-55; 60-62; 63; 64; 104; 131; 177; 221 scoop out, 77 scope of predication, 73; 104; 113; 117; 171 scratcher, 137 seduce, 74

INDEX OF SUBJECTS self-instigation, 47; 65; 67; 77; 82; 137; 173; 209; 224; 229;230;233 sell, 77; 78; 81; 85; 94; 133; 230; 232 seller, 138-39; 140 semantic network, 49-55; 63-64 semantic polygenesis, 216 semasiological variation, 237 sequential scanning, 11; 113 shake, 233 shiver, 234 shiverable, 234 shoot, 57 shudder, 233 Silverstein's hierarchy, 103 sink, 72 sipper, 137 slash, 37; 128 slaughter, 113-15 slaughterable, 83 slay, 111-13 slayable, 83; 226 smash, 36; 236 smasher, 236 smoke (verb), 142 smore, 156-57; 160-61 smother, 156-57; 160-61 smotherable, 85 smoulder, 156-57; 160-61 solidify, 26 solvable, 84 spill, 118 splinter, 36 squash, 37 stab, 23; 57 stage model, 32 starve, 87-89; 225 STARVE verbs, 57; 87-91

starver, 135 steal, 79 stick, 136 stifle, 164-65

stop, 43 stows, 77

strangle, 156; 165-68 stranglehold, 19 stratum, 15 string (verb), 57 stumble, 92 suffocate, 158-59 SUFFOCATE verbs, 56-57; 92-94 doublets, 154-55 etymology, 153-54 metaphorical overlapping, 173-78 overview, 158 round up, 222-24 summary scanning, 11 SUPPLY verbs, 171

switch off, 86 symbolic unit, 9; 13 symbolization, 15 symbolize, 39 Systemic-Functional Grammar, 4; 13-14

Τ taker, 132 tear, 36 thematic participant, 33-35 thicken, 26 thirst, 89-91 throttle, 165-68 throttle an engine, 94-96; 225 throw, 126 tie (verb), 77 torture (verb), 142 touch, 145 trajector, 12; 123; 149; 150; 181; 184 transformations, 15 transitivization, 96; 97; 208-15 translate, 72 typology, 28-30; 45

267

268


u

W

Unaccusative Hypothesis, 16; 45 undryupable, 85 unergative, 45 ungetonable-with, 85 unintentional Actor, 118-19 unkillable, 83 unrelyuponable, 85

wash, 77; 81; 85; 94; 143; 230; 232 watch, 79 with, 171-73 work, 86 worry, 39 WSJ corpus, 18-20

ν variable, 84 verbs of killing (overview), 55-60 voice syncretism, 40; 74; 82; 84; 119; volition, 99-106; 118-19 volitionality see volition.

Y y earn, 57; 90

Ζ zero nominalization, 117 zeugma, 64

Lexical Perspectives on Transitivity and Ergativity: Causative Constructions in English

Perspectives on Classifier Constructions in Sign Languages

Perspectives on Classifier Constructions in Sign Languages

Corpus, Cognition and Causative Constructions (Studies in Corpus Linguistics)

Functional Approaches to Spanish Syntax: Lexical Semantics, Discourse and Transitivity

Stretched Verb Constructions in English

Functional Approaches to Spanish Syntax: Lexical Semantics, Discourse and Transitivity

Particles: On the Syntax of Verb-Particle, Triadic and Causative Constructions (Oxford Studies in Comparative Syntax)

Particles: On the Syntax of Verb-Particle, Triadic and Causative Constructions (Oxford Studies in Comparative Syntax)

Case Valency And Transitivity

Lexical Phonology and the History of English

Subject, Voice and Ergativity

Lexical gestures and Lexical Access

Ergativity in Amazonia

Ergativity (Cambridge Studies in Linguistics)

Case, Valency And Transitivity

Constructions of Colonialism: Perspectives on Eliza Fraser's Shipwreck

Constructions of colonialism: perspectives on Eliza Fraser's shipwreck

Constructions

Lexical Representation and Process

Lexical-Semantic Relations: Theoretical and practical perspectives (Lingvisticæ Investigationes Supplementa)

Morphology and Lexical Semantics

English Adjectives of Comparison: Lexical and Grammaticalized Uses (Topics in English Linguistics)

Lexical representation and process

Constructions and Language Change

Teoria lexical

Lexical Semantics

Perspectives on Death and Dying

Changing Valency: Case Studies in Transitivity

Perspectives On Rehabilitation And Dementia

Lexical Perspectives on Transitivity and Ergativity: Causative Constructions in English

Perspectives on Classifier Constructions in Sign Languages

Perspectives on Classifier Constructions in Sign Languages

Corpus, Cognition and Causative Constructions (Studies in Corpus Linguistics)

Functional Approaches to Spanish Syntax: Lexical Semantics, Discourse and Transitivity

Stretched Verb Constructions in English

Functional Approaches to Spanish Syntax: Lexical Semantics, Discourse and Transitivity

Particles: On the Syntax of Verb-Particle, Triadic and Causative Constructions (Oxford Studies in Comparative Syntax)

Particles: On the Syntax of Verb-Particle, Triadic and Causative Constructions (Oxford Studies in Comparative Syntax)

Case Valency And Transitivity

Lexical Phonology and the History of English

Subject, Voice and Ergativity

Lexical gestures and Lexical Access

Ergativity in Amazonia

Ergativity (Cambridge Studies in Linguistics)

Case, Valency And Transitivity

Constructions of Colonialism: Perspectives on Eliza Fraser's Shipwreck

Constructions of colonialism: perspectives on Eliza Fraser's shipwreck

Constructions

Lexical Representation and Process

Lexical-Semantic Relations: Theoretical and practical perspectives (Lingvisticæ Investigationes Supplementa)

Morphology and Lexical Semantics

English Adjectives of Comparison: Lexical and Grammaticalized Uses (Topics in English Linguistics)

Lexical representation and process

Constructions and Language Change

Teoria lexical

Lexical Semantics

Perspectives on Death and Dying

Changing Valency: Case Studies in Transitivity

Perspectives On Rehabilitation And Dementia

Recommend Documents