Modern Grammars of Case anderson phần 1 pot

4.2 The after-life of ‘deep structure’ 61Part II The Implementation of the Category of Case 6.1 ‘Syntactic/logical’ case forms and localism 1166.1.1 Nominative and genitive 1176.1.2 Dati

Modern Grammars of Case

The past is not dead It is not even past

William Faulkner

This page intentionally left blank

Great Clarendon Street, Oxford OX2 6DP

Oxford University Press is a department of the University of Oxford.

It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide in

Oxford New York

Auckland Cape Town Dar es Salaam Hong Kong Karachi

Kuala Lumpur Madrid Melbourne Mexico City Nairobi

New Delhi Shanghai Taipei Toronto

With offices in

Argentina Austria Brazil Chile Czech Republic France Greece

Guatemala Hungary Italy Japan Poland Portugal Singapore

South Korea Switzerland Thailand Turkey Ukraine Vietnam

Oxford is a registered trade mark of Oxford University Press

in the UK and in certain other countries

Published in the United States

by Oxford University Press Inc., New York

ß John M Anderson 2006

The moral rights of the author have been asserted

Database right Oxford University Press (maker)

First published 2006

All rights reserved No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press,

or as expressly permitted by law, or under terms agreed with the appropriate reprographics rights organization Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above

You must not circulate this book in any other binding or cover

and you must impose the same condition on any acquirer

British Library Cataloguing in Publication Data

Data available

Library of Congress Cataloguing in Publication Data

Data available

Typeset by SPI Publisher Services, Pondicherry, India

Printed in Great Britain

on acid-free paper by

Biddles Ltd., King’s Lynn

ISBN 0-19-929707-x 978-0-19-929707-8

1 3 5 7 9 10 8 6 4 2

Part I The Tradition

2 The Classical Tradition and its Critics 11

2.1.1 The syntax of case and adposition 122.1.2 Grammatical versus local cases 142.1.3 Primary and secondary functions 192.1.4 Conclusion: what is a grammar of case? 222.2 The autonomists and other critics of the tradition 242.2.1 The ‘new grammarians’ 242.2.2 Jespersen versus Hjelmslev on case 272.2.3 Early transformational-generative grammar 29

3.1 The Fillmorean initiative 37

3.1.3 ‘Cases’ and the subject-selection hierarchy 433.1.4 Conclusion and prospect 453.2 The representation of case relations and forms 46

4.2 The after-life of ‘deep structure’ 61

Part II The Implementation of the Category of Case

6.1 ‘Syntactic/logical’ case forms and localism 1166.1.1 Nominative and genitive 1176.1.2 Dative and accusative 119

7 The Variety of Grammatical Relations 149

7.2 Subjecthood and the non-universality of syntax 1517.3 The function of subjects and other grammatical relations 1587.4 The continuum of grammatical relations 1627.5 Ergativity and agentivity 167

8.1 ‘Case’ as a functional category 178

vi Contents

9 The Lexical Structure and Syntax of Functors 220

9.2 Participants and circumstantials 2289.2.1 Circumstantials in ‘case grammar’ 2289.2.2 Apposed circumstantials 2349.2.3 A localist analysis of circumstantials 2359.2.4 Nominals and circumstantials 2429.2.5 Conclusion: circumstantials, incorporation,

9.3 The ineluctability of semantic relations 2459.3.1 The irrelevance of UTAH 2469.3.2 ‘Abstract syntax’ syndrome I: ‘generative semantics’ 2529.3.3 A lexical account of causative constructions 2579.3.4 ‘Abstract syntax’ syndrome II: ‘argument structure’ 267

Part III Case Grammar as a Notional Grammar

10 Groundedness: The Typicality of Case 28110.1 The groundedness of word classes 283

10.1.2 The syntactic consequences of lexical structure 29010.2 The syntactic-categorial structure of words 29410.2.1 Requirements on syntactic categorization 29510.2.2 Parts of speech versus categories 300

10.3.1 Attributive modifiers 307

Contents vii

10.3.2 Noun complements 31010.3.3 Genitival constructions 31610.3.4 Conclusion: apologia 32410.4 Conclusion: ‘notional grammar’ 324

11 Argument-Sharing I: Raising 32711.1 Autonomy and transformations 33111.2 The role of the absolutive 33411.2.1 The status of free absolutive 33411.2.2 The basic syntax of raising: raising with operatives 33711.2.3 Raising with ‘intransitive’ verbs 34011.2.4 Raising with ‘transitive’ verbs 34211.2.5 The category of the infinitive 344

12 Argument-Sharing II: Control 34912.1 The role of the absolutive 34912.1.1 Raising versus control 35212.1.2 Agentive control and the agentivity requirement 35912.1.3 Causatives and control 362

viii Contents

This book addresses a piece of relatively recent history and its continuingconsequences I should acknowledge that it is ‘a personal history’: I am notremote, in any sense, from some of the events of the ‘history’; I am not animpartial historian, and cannot pretend to be one So the ‘history’ not onlysuffers from gaps in my knowledge and understanding—and no doubt in mysympathies; it has also assumed a shape that would almost certainly not havebeen given it by any other narrator Moreover, if I can indulge in moreexplanation of the reasons for the continuing scare quotes around ‘history’,what follows is not a strict chronicle, insofar as what there is of ‘history’ isintermeshed with reinterpretations and reassessments and other after-thoughts concerning the proposals and disputes that form much of the matter

of the book I am primarily concerned with what of the ‘history’ I see asimportant now, not necessarily with how different developments were viewed

at earlier times, though I shall try to document how earlier reactions and reactions have had an effect on this history and on present-day attitudes Butsince the subject of the ‘history’ itself is not temporally remote events, whatseems important, even as it strikes a single person, will doubtless changebefore long To sum up, what is offered here cannot pretend, of course, tosubstitute for direct consultation of the record: it provides only one perspec-tive on the development of the complex of issues that have arisen and arise out

non-of recent concerns with the grammar non-of case

The book grew out of preparations for seminars and lectures to be given atthe Universities of Toulouse II and Bordeaux III, June 2004, one of them atthe conference ‘Journe´es de Linguistique Anglaise’, in Toulouse, 17–18June 2004, organized by the E´quipe de Recherche en Syntaxe et Se´mantique(ERSS) (UMR 5610) The others (Toulouse 15–16 June 2004, Bordeaux 21 June2004) constituted part of the ‘Perpaus’ programme, the Peripatetic Seminar

on Language, Computation and Cognition I am very grateful to thoseresponsible for the organization of these events for, among other things, theopportunity to have been able, in this extended way, to expose to my peerssome of my thoughts on the development of grammars of case Theseheartfelt thanks go particularly to Jacques Durand (ERSS, Toulouse), AnnePrzewowny and Jean Pamie`s (De´partement des E´tudes du Monde Anglo-phone, Toulouse), Claude Mu¨ller (Bordeaux), and Michel Aurnague(University of Pau)

The varied discussions that accompanied the above presentations did much

to contribute to the form and to modify the content of the first five chapters

of this book: considerations of time and compassion ensured that these suffering and stimulating audiences were spared most of what is discussed inthe rest It is invidious to single out particular participants on these occasions,but I must acknowledge the particularly helpful comments and questionsproffered by Christian Bassac, Jacques Durand, Andre´e Morillo, ClaudeMu¨ller, and Jean Pamie`s A revision of these presentations appears in theseries Carnets de grammaire (ERSS, UMR 5610, CNRS and Universite´ deToulouse-Le Mirail) no 15 (2005) That version profited from the commentsand suggestions of Jacques Durand

long-As usual, written versions of (parts of) the book have also benefited fromthe perceptive comments and suggestions of Roger Bo¨hm and Fran Colman,

as also from Jacques Durand’s and Christian Bassac’s continuing interest andstimulus The extent of acknowledgment in the text of the contribution of thefirst of these does not do justice to the extent of his influence on it; and I shall

no doubt again regret not making more of his attempts to save me frommyself This version is also dependent on the comments of two anonymousreaders The volume would not be, without the help and encouragement ofJohn Davey, Consultant Editor Linguistics, Humanities and Social Sciences,OUP

The book is dedicated to John Lyons, who honoured the Toulouse ence with his presence, as did his wife Danielle:

confer-Lampdia  eexont«§ diadvvsoysin allhhloi§ (Plato)

He it was who first accused me of being a ‘localist’; but he is not to blame, anymore than the others mentioned above, for what I have made of it, or theother ideas discussed here

J.M.A.Methoni Messinias, Greece

July2005

x Preface

Conventions and Abbreviations

Examples are numbered consecutively throughout each chapter, (1) to (n).References to and re-presentation of examples in other chapters are preceded

by the chapter number, so that (4.3) is example (3) in Chapter 4; butthe chapter number is omitted with (reference to) examples in the currentchapter Cited words (and lexemes) and word forms are not distinguishedtypographically or otherwise, since it should be clear from the context which

is intended

On grounds of practical economy, the previous work of the present author

is invoked as simply ‘Anderson (date etc.)’, and that of Stephen Anderson as

‘S.R Anderson (date etc.)’

The following abbreviations are used in glosses of examples, where thepractice recommended by the Leipzig glossing rules is followed whereappropriate The rules are available at: http://www.eva.mpg.de/lingua/files/morpheme.html

Abbreviations for semantic relations

Fillmore (1968a) Anderson (1971b/1977) Suggested here

F Factitive loc locative

I Instrumental prt partitive Second order

Certainly, I think one can establish something distinctive about the core ofthe ‘case grammar’ tradition that has evolved over the last forty years; and this

is one of my main aims here But this can be established most transparentlyagainst the background of other work of the same period—and, to someextent more importantly, of the period before Here I follow the recommen-dation of Lyons (1965: 7): ‘Nothing is more helpful in acquiring an under-standing of the principles of modern linguistics than some knowledge of thehistory of the subject.’

As is usual in connection with any scholarly enterprise, the recent(-ish)ideas about ‘case’ that I’m going to examine are often not entirely novel; and it

is important to understand why in some instances we Wnd a continuation anddevelopment of earlier work and in others more drastic revision and rejection.Only thus can we achieve a non-parochial perspective in the evaluation of theadequacy as well as the originality of present-day opinions And, in general,knowledge of the past may at least help us avoid some overgrown gardenpaths

So, though I am focusing on ‘modern grammars of case’, theories primarily

of the twentieth century, work of the preceding decades, which embodiedtraditions going back some centuries, has a role to play in the developmentand evaluation of recent theorizing and its consequences It oVers a baselinefrom which to survey more recent developments and to evaluate the extent to

which they oVer any progress over the tradition, or have failed to avoid itsmistakes This earlier tradition is recognized, at least symbolically, in the title

of one of the earliest publications in ‘case grammar’—Charles Fillmore’s

‘Toward a Modern Theory of Case’, of 1965 The title also encapsulates theambivalence of the term ‘case’, as denoting either the relations (semantic orgrammatical) expressed by morphological case or that morphological means

of expression itself

In modern work on ‘case’, in either sense, much of the acknowledgment ofthe contribution of earlier work is (as in this title) implicit only, thoughFillmore (1968a), for instance, does oVer a brief critique of the practice ofsome previous grammars of case But, as anticipated, I shall try to make thisdebt a bit more overt as we proceed—and, indeed, from the very beginning

Of course, even this is limited in the present work by the space ately available for such ‘contextualization’ However, Chapter 2, ‘The ClassicalTradition and its Critics’, endeavours to provide some background to thedevelopments stemming from the third quarter of the last century whoseevolution we are primarily concerned with here, as well as to establish theextent to which ‘case grammar’, compared with other modern treatments ofcase, maintains traditional ideas of case and its centrality in the grammar

proportion-In what immediately follows that chapter, what I see as the main conceptsthat emerged as a rough consensus from the earliest embodiments of ‘casegrammar’ are our immediate concern This consensus takes over the trad-itional notion that ‘case forms’ express both semantic relations (such as

‘agent’ and ‘location’) and grammatical relations (such as ‘subject’) Theseagreed concepts will occupy us in Chapters 3–5 Thereafter we shall beconcerned with ‘unWnished business’ from these early years

In the Wrst place there are central issues which were not resolved at thattime, one of which, the question of the set of semantic ‘cases’, or semanticrelations, already emerges as such in Chapter 5 The latter part of that chapter

is devoted to one attempt to resolve the question of the identity of ‘cases’ and

of ‘case’, an undertaking whose origins are rather ancient, namely the called ‘localist theory of case’, whose early implementation in a variety of ‘casegrammar’ is discussed there

so-The scare quotes around ‘case grammar’ are a reminder that this approach

is only one variety of a grammar of case Those around ‘cases’ and ‘case’recognize that it has been acknowledged for some time that the relationsexpressed by morphological case can be expressed in other ways, notably byadpositions and position ‘Case’ refers to these common relations; and mor-phological case is only one kind of ‘case form’, one way of expressing ‘caserelations’, or simply ‘case’

2 Modern Grammars of Case

With Chapter 6, which looks at more recent attempts to implement thelocalist hypothesis, we move on to more recent developments in ‘localist casegrammar’ This chapter thus initiates discussion of ideas that emerged afterthe earliest period of ‘case grammar’ We can conveniently locate the end ofthis period in the late 1970s, the time of Fillmore’s partial ‘retraction’ (1977)and of ‘defences’ of the ‘case grammar’ hypothesis against early criticismssuch as Anderson (1977).

Chapter 7 pursues questions to do with the articulation of the relationshipbetween ‘case’ relations and ‘case forms’, the morphological, lexical, andpositional signals of the ‘cases’, including the ways in which the ‘case forms’neutralize expression of the semantic or ‘case’ relations It looks at argumentsthat the patterns of neutralization, such as subject formation, are not univer-sal, though the variant possibilities show similarities, including functionalmotivation

Chapter 8 focuses on attempts to resolve the question of the categoriality of

‘case’: what kind of category do the ‘cases’ belong to? It is a type of categorythat, as I have indicated, can be manifested in various ways, as e.g prep-osition, inXectionally, or by position How is this to be accommodated,and how is the co-presence of these diVerent kinds of manifestation in thesame language system to be articulated? This concerns the status of ‘func-tional’ categories Chapter 9 then looks in more detail at the lexical structureand the syntax of the category of ‘case’, the ‘functor’, in the terminologyadopted there

The concluding chapters of the book concern a slightly diVerent kind ofunWnished business In various ways the ‘cases’ could be seen as slightlyanomalous within the framework of assumptions that determined the shapeand substance of the grammar in which they were initially embedded Thereare at least two important aspects to this

As I shall discuss in Chapter 10, the ‘cases’ are clearly grounded in semanticsubstance: they are identiWed semantically and their semantics determinestheir basic distribution The implementation of this identiWcation has beencontroversial (and remains so); but it has generally been thought to be anappropriate pursuit, even by those who would reduce ‘case’ to a conjunction

of other categories And even in Starosta’s austerely autonomous (fromsemantics) development of ‘case grammar’, the ‘case relations’ are regarded

as ‘still meaningful, but in a quite abstract and general way’ (1988: 123) Theextent to which other syntactic categories are similarly grounded was one ofthe unresolved issues of early ‘case grammar’ (and this was matched by similarcontroversy in other approaches to grammar in which grounding was notsimply denied)

Prologue 3

However, the consequences of a decision in this area go far beyond theconWnes of the original ‘case grammar’ programme It is only rather morerecently that there has been given any recognition to the conclusion that ‘casegrammar’ is simply a sub-theory of a general ‘notional’, or ‘ontologicallybased’ grammar, and that an assumption of autonomy for syntax is no lessinjurious elsewhere in the grammar than it is in relation to the ‘cases’ Chapter

10 looks at proposals to apply ‘groundedness’ in the study of syntacticcategories in general Just as phonology is regarded by many phonologists

as ‘grounded’ in phonetic substance, so too syntactic categories and ships are ‘grounded’ in meaning This is one aspect of a more general variety

relation-of unWnished business

In the second place, other questions arise from the fact that, with minordepartures consequent upon recognition of the centrality of ‘cases’, Fillmore’s(1968a) proposals are embedded in a standard transformational grammar ofthe time; and this emerges as a particularly salient issue in a ‘case grammar’.From diVerent points of view, there developed in subsequent work a convic-tion that this was undesirable: the transformational apparatus is not onlyundesirable in itself, but it is also especially inappropriate, and unnecessary, in

a ‘case grammar’ In the Wnal chapters of the book I shall look at how some ofthe properties that have accrued to ‘case grammar’ have been said to rendersuperXuous any appeal to transformations and their equally unpalatableconcomitant, ‘empty categories’

What emerges overall from these more recent developments and theirongoing continuation is an understanding of the extent to which an appeal

to the autonomy of the ‘computational system’ has grossly distorted linguists’conception of the relationship between meaning and grammar It has more-over led to a perverse characterization of what counts as ‘linguistic creativity’,reducing it to a by-product of the ability to compute recursive routinized(meaning-free) formulae For Foley and van Valin, for instance, ‘linguisticcreativity’ is ‘the ability of native speakers to produce and understand an (inprinciple) inWnite number of sentences’ (1984: 319) This lays emphasis on thecomputational capacity underlying this ability to cope with an ‘inWnity’ ofsentences But creativity in language, on any normal understanding, involves,rather, the capacity to formulate representations for newly perceived ‘scenes’and to decode them, possibly in novel situations; and it includes lexical as well

as (and probably more so than) syntactic capacities This capacity depends on

an understanding of Wgurativeness, which, despite the relative routinization,

or institutionalization, involved in lexicalization and grammaticalization, isbasic to the structure and development of both lexical and grammaticalsystems

4 Modern Grammars of Case

The notion of governed creativity’ (Chomsky 1976), versus breaking creativity’, involves a misapprehension: rules cannot govern

‘creativity’; rather, they may help to enable creativity, or to provoke to breaking creativity’ ‘Rule-governed creativity’ is a misnomer We have a termalready for ‘creativity’ that is said to be ‘rule-governed’: it is usually called

‘rule-‘(recursive) productivity’ Such productivity has a minor contribution to make

to creativity, but it should not be identiWed with it or considered basic to it

It is also misleading to describe ‘literary’ creativity as ‘rule-breaking’;typically it is ‘rule-extending’ or ‘rule-making’ (for example Thorne 1965;1969; and other references in Thorne 1970) When, for instance, to take asimple example, Peter Carey writes—or, rather, one of his characters says—

‘She grew me up’ (Jack Maggs, ch 26), he is ‘extending’ the lexical incidence ofcausativization by conversion (cf the lexical causative Bring up) It is obvioustoo that such creativity is not conWned to ‘literature’; it is basic to our capacity

to use language to express our perceptions and to interpret the expressions

of others

As implied by the preceding chapter descriptions and groupings, I havedivided the set of chapters which follow this Prologue into three parts,followed by an Epilogue Part I, ‘The Tradition’, discusses, against the earlierbackground, the evolution of grammars of case in the twentieth century,particularly its third quarter, and particularly the early development of theapproach that came to be called ‘case grammar’ This part comprises Chapters2–5, terminating in the chapter on the identity of semantic relations

The boundary between Parts I and II cuts across a grouping implied in thedescription of the chapters given above: both the latter part of Chapter 5 andChapter 6 are concerned with localism As I’ve suggested already, one reasonfor this is that all of the material in Parts II and III concerns developments thatare for the most part later than the work discussed in Part I The proposedincorporation of the ‘localist hypothesis’ into ‘case grammar’ occurred quiteearly in the evolution of the latter; and, though its status was not generallyagreed on, it was adopted rather early in some form in a variety of approachesstemming from ‘case grammar’ The developments presented in Chapter 6,however, though based on the ‘localist hypothesis’, belong to a much laterperiod, and are indeed partly original to this volume There is a chronologicalmotivation for the division

Another reason for the proposed division between Parts I and II is that thework discussed from that point on focuses on attempts to articulate moreexplicitly the basic ideas discussed in Part I and their consequences Thediscussion from this point on also shows a further admitted narrowing

of ‘scholarly focus’, in its concentration on developments in the localist

Prologue 5

interpretation of ‘case grammar’ Nevertheless, there is also an attempt there

to give attention to diVering viewpoints, both in work which regards itself as

‘case grammar’ and in studies that do not, as illustrated by concern with thestatus if any of ‘macro-roles’ and ‘abstract syntax’ (Chapter 9)

Part II, ‘The Implementation of the Category of Case’, comprises thosechapters, 6–9, that seek to establish the categorial character of ‘case’ in thewide sense Part II is divided from Part III not for reasons of chronology, sincethe proposals discussed in the two parts developed in parallel, and not entirelyindependently; rather, as anticipated in the brief descriptions of the individ-ual chapters, the division reXects again a diVerence in focus, but in this case infocus within the grammar

While Part II is concerned very much with the category of ‘case’, and otherfunctional categories, the chapters that follow involve the recognition that

‘case’ is typical of syntactic categories in being semantically grounded: Part III

is called ‘Case Grammar as a Notional Grammar’ This is introduced inChapter 10; Chapters 11 and 12 concern work that examines more explicitlythe role of ‘case’ in the syntax and morphology of such a ‘notional grammar’,and particularly its part in eliminating appeal to syntactic transformationsand other syntactic paraphernalia, in favour of simple projections from arichly structured but formally parsimonious lexicon

The book closes with an Epilogue which tries to draw together the mainresults, as I see them, of the various developments in grammars of case in thechapters which precede it, as well as to point to some extensions and furtherconsequences of the main traditions which can be described as grammars ofcase The history oVered here is too personal, and too (re)interpretative (oftenwith the beneWt, or handicap, of hindsight) to count as historiography; it alsotranscends the historiographical in oVering novel analyses of many of thephenomena considered—not just in the Epilogue but also, for instance, in thediscussion of localism in Chapter 6 It is a history, from my viewpoint, ofcertain ideas whose development is as informative as the form they take at anyone period; the present has only a minor privilege in this respect Thisdevelopmental orientation means that analyses are presented as evolvingrather than as having assumed some ‘Wnal’ form, so that, for example, thetreatment of passives or causatives is recurrently modiWed in the light ofconceptual shifts This orientation also underlies the alternation in the textbetween more panoramic views of general developments and explicit anddetailed concern with the motivations and consequences of these as exem-pliWed by particular analyses

A historical perspective keeps before us the contingency of our theoreticalassumptions, and their unsuitability for constituting dogma The result of the

6 Modern Grammars of Case

approach adopted here is that, as well as not being properly ical, the presentation departs from the usual formula of: exposition (or oftensimply assumption) of the theoretical framework; consideration of previousresearch within that general framework on a particular area of variable scope;(re-)application to the area of the framework, or some limited revision of it.This scarcely seemed to be appropriate to the (re-)evaluative goals of thepresent enterprise It is also salutary, I suggest, for us to give up now andagain the pretence that linguistic research can only be pursued as if only what isfamiliar today is what is most relevant Every epilogue is also a prologue.

historiograph-Prologue 7

This page intentionally left blank

Part I

The Tradition