class="bi x0 y0 w0 h1"
Modern Grammars of Case
The past is not dead. It is not even past.
William Faulkner
This page intentionally left blank
Modern Grammars
of Case
A Retrospective
JOHN M. ANDERSON
1
3
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

2.2 The autonomists and other critics of the tradition 24
2.2.1 The ‘new grammarians’ 24
2.2.2 Jespersen versus Hjelmslev on case 27
2.2.3 Early transformational-generative grammar 29
2.3 Conclusion 35
3 Early Case Grammar 36
3.1 The Fillmorean initiative 37
3.1.1 ‘Cases’ and grammar 38
3.1.2 Linearity 41
3.1.3 ‘Cases’ and the subject-selection hierarchy 43
3.1.4
Conclusion and prospect 45
3.2 The representation of case relations and forms 46
3.2.1 Dependency 46
3.2.2 The categorial identity of case and preposition:
a functional category 48
3.2.3 ‘Case’ and position 51
3.2.4 Conclusion 52
3.3 Conclusion 53
4 Case Grammar and the Demise of Deep Structure 56
4.1 ‘Deep structure’ and the place of holisticness 57
4.2 The after-life of ‘deep structure’ 61
4.2.1 ‘Unaccusativity’ 62
4.2.2 Lexical evidence 66
4.2.3 Raising 69
4.2.4 Conclusion 72
4.3 Excursus on the tortuous history of ‘thematic relations’ 74
4.4 Conclusion: where we have reached 76
5 The Identity of Semantic Relations 79
5.1 Distributional criteria for particular ‘cases’ 80

8.2.2 Determination 184
8.2.3 Conclusion 186
8.3 Kuryłowicz’s problem 187
8.3.1 A solution: the Latin accusative 188
8.3.2 The Latin case system, and an alternative solution 196
8.3.3 Case in English 208
8.3.4 Conclusion: functors and lexical structure 211
8.4 Complex cases: Hjelmslev on Tabasaran 212
8.5 Conclusion and consequences 218
9 The Lexical Structure and Syntax of Functors 220
9.1 ‘Macroroles’ 220
9.2 Participants and circumstantials 228
9.2.1 Circumstantials in ‘case grammar’ 228
9.2.2 Apposed circumstantials 234
9.2.3
A localist analysis of circumstantials 235
9.2.4 Nominals and circumstantials 242
9.2.5 Conclusion: circumstantials, incorporation,
and absorption 244
9.3 The ineluctability of semantic relations 245
9.3.1 The irrelevance of UTAH 246
9.3.2 ‘Abstract syntax’ syndrome I: ‘generative semantics’ 252
9.3.3 A lexical account of causative constructions 257
9.3.4 ‘Abstract syntax’ syndrome II: ‘argument structure’ 267
9.4 Conclusion 273
Part III Case Grammar as a Notional Grammar
10 Groundedness: The Typicality of Case 281
10.1 The groundedness of word classes 283
10.1.1 Verbs and nouns 285
10.1.2 The syntactic consequences of lexical structure 290

12.3 Conclusion 379
13 Epilogue: Case, Notionalism, Creativity, and the Lexicon 381
13.1 Retrospect 381
13.2 Lexical structure 385
13.2.1 Complex predicators 386
13.2.2 Argument-linking 389
13.2.3 Constraints on valency 399
13.2.4 Lexical structure and morphology 404
13.2.5 Absorption, incorporation, and ‘constructions’ 407
13.3 Creativity and notionalism 412
References
419
Index 441
viii Contents
Preface
This book addresses a piece of relatively recent history and its continuing
consequences. I should acknowledge that it is ‘a personal history’: I am not
remote, in any sense, from some of the events of the ‘history’; I am not an
impartial historian, and cannot pretend to be one. So the ‘history’ not only
suffers from gaps in my knowledge and understanding—and no doubt in my
sympathies; it has also assumed a shape that would almost certainly not have
been given it by any other narrator. Moreover, if I can indulge in more
explanation of the reasons for the continuing scare quotes around ‘history’,
what follows is not a strict chronicle, insofar as what there is of ‘history’ is
intermeshed 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 w ith what of the ‘history’ I see as
important now, not necessarily with how different developments were viewed
at earlier times, though I shall try to document how earlier reactions and non-
reactions have had an effect on this history and on present-day attitudes. But

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 long-
suffering and stimulating audiences were spared most of what is discussed in
the rest. It is invidious to single out particular participants on these occasions,
but I must acknowledge the particularly helpful comments and questions
proffered by Christian Bassac, Jacques Durand, Andre
´
e Morillo, Claude
Mu
¨
ller, and Jean Pamie
`
s. A revision of these presentations appears in the
series Carnets de grammaire (ERSS, UMR 5610, CNRS and Universite
´
de
Toulouse-Le Mirail) no. 15 (2005). That version profited from the comments
and suggestions of Jacques Durand.
As usual, written versions of (parts of) the book have also benefited from
the perceptive comments and suggestions of Roger Bo
¨
hm and Fran Colman,
as also from Jacques Durand’s and Christian Bassac’s continuing interest and
stimulus. The extent of acknowledgment in the text of the contribution of the
first of these does not do justice to the extent of his influence on it; and I shall

The following abbreviations are used in glosses of examples, where the
practice recommended by the Leipzig glossing rules is followed where
appropriate. The rules are available at: />morpheme.html
Abbreviations in glosses
A actor (Tagalog)
ABL ablative
ABS absolutive
ACC accusative
ACT active (Malagasy)
ADS adessive
AGR agreement
AGT agent
ALL allative
ANTIP antipassive
ASP aspect
AT actor-topic
CAUS causative
CIRC circumstantial (Malagasy)
CL class marker
D direction (Tagalog)
DAT dative
DEF deWnite
DISTR distributive
DT direction-Topic
ERG ergative
ESS essive
F feminine
FUT future
G goal (Malagasy, Tagalog)
GEN genitive

F Factitive loc locative
I Instrumental prt partitive Second order
L Locative goal
O Objective src source
T Time
Fillmore elsewhere (only those alluded to here, and ignoring mere terminological vari-
ation)
E Experiencer
G Goal
S Source
P Path
Other category abbreviations
A adjective
C comparator
D determinative
dim dimensional
N noun
N referentiable
P predicative
pass passive
pat patient
prog progressive
T Wniteness
Vverb
/ takes as a complement
\ modiWes
xii Conventions and Abbreviations
1
Prologue
By my title I’ve described the area that I want to look at here as ‘modern

which they oVer any progress over the tradition, or have failed to avoid its
mistakes. 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 the
ambivalence of the term ‘case’, as denoting either the relations (semantic or
grammatical) expressed by morphological case or that morphological means
of expression itself.
In modern work on ‘case’, in either sense, much of the acknowledgment of
the contribution of earlier work is (as in this title) implicit only, though
Fillmore (1968a), for instance, does oVer a brief critique of the practice of
some previous grammars of case. But, as anticipated, I shall try to make this
debt 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 proportion-
ately available for such ‘contextualization’. However, Chapter 2, ‘The Classical
Tradition and its Critics’, endeavours to provide some background to the
developments stemming from the third quarter of the last century whose
evolution we are primarily concerned with here, as well as to establish the
extent to which ‘case grammar’, compared with other modern treatments of
case, maintains traditional ideas of case and its centrality in the grammar.
In what immediately follows that chapter, what I see as the main concepts
that emerged as a rough consensus from the earliest embodiments of ‘case
grammar’ 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’). These
agreed concepts will occupy us in Chapters 3–5. Thereafter we shall be
concerned with ‘unWnished business’ from these early years.
In the Wrst place there are central issues which were not resolved at that
time, one of which, the question of the set of semantic ‘cases’, or semantic
relations, 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

same language system to be articulated? This concerns the status of ‘func-
tional’ categories. Chapter 9 then looks in more detail at the lexical structure
and the syntax of the category of ‘case’, the ‘functor’, in the terminology
adopted there.
The concluding chapters of the book concern a slightly diVerent kind of
unWnished business. In various ways the ‘cases’ could be seen as slightly
anomalous within the framework of assumptions that determined the shape
and substance of the grammar in which they were initially embedded. There
are at least two important aspects to this.
As I shall discuss in Chapter 10, the ‘cases’ are clearly grounded in semantic
substance: they are identiWed semantically and their semantics determines
their basic distribution. The implementation of this identiWcation has been
controversial (and remains so); but it has generally been thought to be an
appropriate pursuit, even by those who would reduce ‘case’ to a conjunction
of other categories. And even in Starosta’s austerely autonomous (from
semantics) development of ‘case grammar’, the ‘case relations’ are regarded
as ‘still meaningful, but in a quite abstract and general way’ (1988: 123). The
extent to which other syntactic categories are similarly grounded was one of
the unresolved issues of early ‘case grammar’ (and this was matched by similar
controversy in other approaches to grammar in which grounding was not
simply denied).
Prologue 3
However, the consequences of a decision in this area go far beyond the
conWnes of the original ‘case grammar’ programme. It is only rather more
recently that there has been given any recognition to the conclusion that ‘case
grammar’ is simply a sub-theory of a general ‘notional’, or ‘ontologically
based’ grammar, and that an assumption of autonomy for syntax is no less
injurious elsewhere in the grammar than it is in relation to the ‘cases’. Chapter
10 looks at proposals to apply ‘groundedness’ in the study of syntactic
categories in general. Just as phonology is regarded by many phonologists

basic to the structure and development of both lexical and grammatical
systems.
4 Modern Grammars of Case
The notion of ‘rule-governed creativity’ (Chomsky 1976), versus ‘rule-
breaking creativity’, involves a misapprehension: rules cannot govern
‘creativity’; rather, they may help to enable creativity, or to provoke to ‘rule-
breaking creativit y’. ‘Rule-governed creativity’ is a misnomer. We have a term
already for ‘creativity’ that is said to be ‘rule-governed’: it is usually called
‘(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 a
simple 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 of
causativization by conversion (cf. the lexical causative Bring up). It is obvious
too 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 have
divided the set of chapters which follow this Prologue into three parts,
followed by an Epilogue. Part I, ‘The Tradition’, discusses, against the earlier
background, the evolution of grammars of case in the twentieth century,
particularly its third quarter, and particularly the early development of the
approach that came to be called ‘case grammar’. This part comprises Chapters
2–5, terminating in the chapter on the identity of semantic relations.
The boundary between Parts I and II cuts across a grouping implied in the
description of the chapters given above: both the latter part of Chapter 5 and
Chapter 6 are concerned with localism. As I’ve suggested already, one reason
for this is that all of the material in Parts II and III concerns developments that

the role of ‘case’ in the syntax and morphology of such a ‘notional grammar’,
and particularly its part in eliminating appeal to syntactic transformations
and other syntactic paraphernalia, in favour of simple projections from a
richly structured but formally parsimonious lexicon.
The book closes with an Epilogue which tries to draw together the main
results, as I see them, of the various developments in grammars of case in the
chapters which precede it, as well as to point to some extensions and further
consequences of the main traditions which can be described as grammars of
case. The history oVered here is too personal, and too (re)interpretative (often
with the beneWt, or handicap, of hindsight) to count as historiography; it also
transcends the historiographical in oVering novel analyses of many of the
phenomena considered—not just in the Epilogue but also, for instance, in the
discussion of localism in Chapter 6. It is a history, from my viewpoint, of
certain ideas whose development is as informative as the form they take at any
one period; the present has only a minor privilege in this respect. This
developmental orientation means that analyses are presented as evolving
rather than as having assumed some ‘Wnal’ form, so that, for example, the
treatment of passives or causatives is recurrently modiWed in the light of
conceptual shifts. This orientation also underlies the alternation in the text
between more panoramic views of general developments and explicit and
detailed concern with the motivations and consequences of these as exem-
pliWed by particular analyses.
A historical perspective keeps before us the contingency of our theoretical
assumptions, 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 historiograph-
ical, the presentation departs from the usual formula of: exposition (or often
simply assumption) of the theoretical framework; consideration of previous
research 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.

subsequent work in the less philosophical strands of the tradition—except
perhaps negatively: not gender, not number. Rather, we get recognition of a
distinction between two kinds of cases: the casus rectus, the nominative, which
marks the subject of the Wnite verb, and the oblique cases, which at least in
some uses signal a semantic relation to the verb, as illustrated in (1):
(1) Missı¯le
¯
ga
¯
tı¯ Athe
¯
na
¯
s sunt
sent envoys:
NOM Athens:ACC are
(‘Envoys were sent to Athens’)
(Gildersleeve and Lodge 1968: 214). Here the accusative marks the spatial goal
of the movement signalled by the verb. The nature of this alleged distinction
between the casus rectus and the others, and variants of such a distinction,
underlie much of the debate within modern grammars of case.
Not so much debated of late has been the problematical status of the vocative
(see, however, Hjelmslev (1935/7) and Mel’c
ˇ
uk (1986), who reject it as a case).
As observed by the ancients, the vocative seems to belong paradigmatically
with the cases, as illustrated by the (modern) Greek paradigm in (2a), but
functionally has little in common with them, as illustrated by (2b):
(2)a.fı
´

Table 2.1 Varro’s Latin word classes
InXected for Case
þÀ
À nouns conjunctions etc.
with
Tense
forms
þ participles verbs
12 Modern Grammars of Case