LINGUIST List 12.2582
Tue Oct 16 2001
Review: Butt & Holloway King, Argument Realization
Editor for this issue: Terence Langendoen <terrylinguistlist.org>
What follows is another discussion note contributed to our Book Discussion
Forum. We expect these discussions to be informal and interactive; and
the author of the book discussed is cordially invited to join in.
If you are interested in leading a book discussion, look for books
announced on LINGUIST as "available for discussion." (This means that
the publisher has sent us a review copy.) Then contact Simin Karimi at
simin
linguistlist.org or Terry Langendoen at terry
linguistlist.org.
Directory
Simon Musgrave, Review of Butt & Holloway King, Argument Realization
Message 1: Review of Butt & Holloway King, Argument Realization
Date: Tue, 16 Oct 2001 10:27:46 +0200
From: Simon Musgrave <S.Musgravelet.leidenuniv.nl>
Subject: Review of Butt & Holloway King, Argument Realization
Butt, Miriam & Tracy Holloway King, eds. (2000) Argument
Realization. CSLI Publications, x+244pp, paperback ISBN
1-57586-266-2, $25.00, hardback ISBN 1-57586-265-4, Studies in
Constraint-Based Lexicalism
Simon Musgrave, Spinoza Project: Lexicon & Syntax
University of Leiden Centre for Linguistics
[An announcement of this book can be found at
http://linguistlist.org/issues/12/12-703.html#1 --reviewer]
This volume is a collection of seven papers dealing in
various ways with the issue of how the semantic arguments
of predicates are realised in the expressions of language.
An introduction by the editors, a subject index and a name
index are also included. All the papers are written within
the theoretical framework of Lexical-Functional Grammar
(LFG)(Bresnan 2001). This theory has three types of
structure which are relevant to the issues with which the
papers deal. Argument structure (a-structure) models the
relations of predicates and arguments independent of any
syntactic encoding. Functional structure (f-structure)
models the relations of predicates and grammatical
functions such as subject and object. And constituent
structure (c-structure) models the surface syntactic form
of linguistic expressions. Correspondence principles
constrain the mapping between structures, but the
information encoded in each structure is distinct. The
questions addressed in this book are about the nature of
these correspondence principles and how they work in
particular languages.
In their introduction, the editors divide the papers into
three groups, those dealing with the issue of the division
of labour between morphology and syntax, those dealing with
complex predicates, and those dealing with linking theory
(Carter 1976). In the terms of the sketch of the
theoretical framework just given, the first group of papers
deal with the correspondence between f-structure and c-
structure and the third group deal with the correspondence
between a-structure and f-structure. The analysis of
complex predicates raises problems for both sets of
correspondence principles: in the first case, the issue of
the shared properties of morphological and periphrastic
complex predications, and in the second case , the issue of
how complex events can be coded as one set of predicate-
argument relations. This grouping is logical, but one
important interrelation is not discussed by the editors.
The papers by Sadler and Laczk� both deal with the status
of semantic arguments within NP, and both analyses
crucially depend on the possessor function.
B�rjars and Vincent (Multiple Case and the 'Wimpishness' of
Morphology - B&V) argue that properties of the
morphological system of a language can constrain the
realisation of syntactic features. The crucial data come
from situations in which a certain NP or element of an NP
can be analysed as having two case features assigned to it.
In some languages, both features can be expressed (case-
stacking or Suffixaufnahme), and in other languages only
one feature is expressed, but this may not be the expected
one (what B&V call over-riding). B&V point out the
typologically interesting fact that there are apparently no
languages with fusional morphology and case-stacking, and
no languages with agglutinative morphology and over-riding.
Therefore, the surface expression of similar feature
structures in different languages is constrained only by
properties internal to the morphological component of the
language.
Nordlinger (Australian Case Systems: Towards a Constructive
Solution - N) also deals with case-stacking in her paper,
which is in effect an introductory presentation of the
material covered in detail in Nordlinger (1998). N's
analysis reverses the usual assumption as to how case-
marking works. Rather than the case morphology being the
reflex of a syntactic feature assigned to a dependent, the
case marker is treated as a feature of the dependent which
constructs the wider environment in which it occurs. Thus,
in a simple case, a NP with accusative case carries the
information that it will be the value of the attribute OBJ
in some larger f-structure (assuming NOM-ACC alignment).
This is handled formally by inside-out function assignment
(Halvorsen & Kaplan 1988). A Principle of Morphological
Composition allows this process to extend to the case-
stacking examples. This principle guarantees that the
information contributed by a morpheme will be unified with
the outermost f-structure constructed by the stem to which
it attaches. The effect of the principle is vacuous except
in those cases where inside-out function assignment is
specified.
Sadler's paper (Noun Phrase Structure in Welsh - S)
addresses the problem of word order in Welsh NPs. The head
is usually initial in such phrases, but a possessor appears
closer to the head than apparent PP complements. This order
has been accounted for by analysing the position of the
head as the result of movement from within NP to some
dominating functional projection. S argues in detail
against one version of this analysis, that of Rouveret
(1994), although the general line of argumentation could be
used against other similar proposals. The alternative
analysis for which S argues is that Welsh nominals, and
probably nominals in all Gaelic languages, do not project
any argument structure except a possessor slot. This is not
to say that deverbal nouns may not have semantic arguments,
but syntactically they must always be realised as
possessors, or as adjuncts which are adjoined to NP. This
analysis accounts for the fact that possessors and
adjectives appear between the head and its apparent PP
complements, and also for some co-ordination facts which
are not easily handled in the head-raising analysis.
Broadwell's paper (Choctaw Directionals and the Syntax of
Complex Predication - B) is an analysis of directional
particles in the American language Choctaw. A similar
function, that of specifying the direction of motion, is
served by verbal morphology in some other languages, such
as Oneida. B analyses both types as complex predications,
the difference between the two being parallel to that
between morphological and periphrastic causatives. The
directional particles normally immediately precede the
verb, but B offers convincing arguments that they are
separate words and that they are not members of the rather
limited class of adverbs in Choctaw. The remainder of B's
paper is devoted to discussing the semantic condition which
governs the use of directional particles. B argues that if
the conceptual representation of a verb includes the
element GO, then a directional particle is compatible with
it but not otherwise. This claim has the interesting
consequence that some verbs which are not obviously motion
verbs must nevertheless, at least for Choctaw, have the
motion component GO included in their conceptual
representation.
Matsumoto (Crosslinguistic Parameterization of Causative
Predicates - M), using data from Japanese, argues that the
typology of causatives is more complex than is assumed by,
for example, Alsina (1997 and elsewhere). M presents data
which suggests that the distinction between two types of
causative which is crucial to Alsina's proposal is a
semantic one and that the arguments which are shared can be
fused in various ways. This allows for the possibility of
causatives which are monoclausal at both a-structure and at
f-structure and M argues that this possibility occurs in
Japanese. Given this evidence, M proposes that a typology
of causatives must take two factors into account: the
semantic distinction noted by Alsina, and the complexity of
the structures involved. Of the six logical possibilities,
M notes that two are not represented in the languages he
has examined, and he argues that these gaps are principled
and not accidental.
L�drup (Underspecification in Lexical Mapping Theory -L)
examines intransitive verbs in Norwegian and presents
evidence that the distinction between unergative and
unaccusative verbs breaks down in some environments.
Existential sentences are predicted only to be possible
with unaccusative verbs, but Norwegian allows unergative
verbs in this clause type. On the other hand, it is well
known that unaccusatives can appear freely with
resultatives while unergatives require a fake reflexive
(Simpson 1983, Levin and Rapaport-Hovav 1995), but in
Norwegian, unaccusatives also can have fake reflexives in
resultative clauses. L argues that these data can be
handled in LFG by underspecification of features in Lexical
Mapping Theory (LMT), the LFG version of linking theory. L
suggests that the basic intuition to be captured is that
agents and themes do not have their prototypical properties
in the absence of the other role.
Lazck�'s paper (Derived Nominals, Possessors, and Lexical
Mapping Theory - La) is a discussion of the properties of
deverbal nominals in Hungarian, focussing on their
argument-taking properties and the status of the possessor
constituent. He demonstrates that the Hungarian NP has only
one argument position available, the possessor slot, and
that this position is filled for event nominalisations on
an ergative basis. That is, the subject of an intransitive
verb will become possessor in the nominalisation, but the
object of a transitive verb will fill this slot. The
subject of a transitive verb can only be realised in a
postpositional phrase. Mapping to the clausal function
subject is reserved for the most thematically prominent
argument (in the normal case), but in Hungarian NPs, the
less thematically prominent argument maps to possessor. La
proposes a reversal of the association principles which LMT
applies for clausal argruments, a stipulative solution.
However, in a footnote, he offers a more convincing
account. The patient-like argument cannot be realised as an
oblique due to a feature clash, while the agent-like
argument can. Therefore, it is only necessary to assume
some principle such as full interpretation to account for
the observed pattern. The final section of La's paper
discusses various properties of the possessive morpheme in
Hungarian.
The majority of these papers were originally presented at
conferences organised by the LFG community, and the first
observation to make in evaluating the volume is that it
reflects an eclectic and active research community. At the
level of presentation, the editing is generally good. A few
misprints remain, the section numbering in Nordlinger's
paper goes awry, and Laczk�'s prose is occasionally
obscure, but the text is readable. However, at the level of
content I am did feel at times that the editors might have
been more demanding, and I will discuss two examples of
this briefly.
Broadwell's paper offers very little argumentation that the
phenomena being discussed are examples of complex
predication. Having shown that in some languages the
relevant meanings are expressed morphologically, and in
some other languages they are expressed syntactically,
Broadwell then assumes that complex predication is
involved. No details are given of how the argument-taking
properties of the verbs change when they combine with
directional particles, nor data as to whether the
directional meaning falls within the scope of negation or
of tense and aspect. The data presented gave me the
impression that an analysis that treated directional
particles as a type of verbal modifier would be just as
valid as the analysis which appeals to complex predication.
The bulk of the paper, which deals with the semantic
conditions constraining the use of directionals would be
unchanged under the alternative analysis, and would retain
its interest.
In L�drup's paper, part of the argument relating to
resultatives depends on the claim that clauses with fake
reflexives must have the same structure whether the verb is
unaccusative or unergative. The unergative verb possibility
is, not surprisingly, taken to be unmarked, the subject of
such a verb is assigned the intrinsic feature
[-o(bjective)] by LMT. Therefore, L�drup argues that in a
clause with an unaccusative verb the subject must also
have the intrinsic feature [-o], against the assumptions of
LMT. But this argument occurs in the paper after it has
already been argued that the subject of unergative verbs
can have the intrinsic feature [-r(estricted)] rather than
[-o]; this assumption is needed to account for the
existential sentence data. This leaves unclear the status
of the argument just sketched: if unergative verbs can have
a [-r] subject in existential sentences, need it be assumed
that they always have a [-o] subject elsewhere?
Additionally, all L�drup's arguments go to show that the
distinction between unergatives and unaccusatives is not
relevant in the grammar of Norwegian. Nevertheless, the
paper is written as though the distinction is valid. It
would have been useful if evidence had been presented
showing where the distinction has consequences which are
not undermined by L�drup's account. As it stands, the paper
gives the impression that the terms 'unergative' and
'unaccusative' are used only to refer to the semantic
character of the various verbs. This point is of more
general theoretical interest within LFG. The capacity of
LMT to provide a natural account of unaccusativity and
resultatives is taken as evidence in favour of the approach
in e.g. Bresnan 2001. As presented, the Norwegian data
weaken this argument, but if it is the case that the
unergative/unaccusative distinction is not relevant in
Norwegian syntax, then the more general argument is not
affected.
These reservations should not discourage readers who are
not LFG devotees from sampling this volume. The papers by
B�rjars & Vincent and by Matsumoto present findings which
will be of interest to many and which are not theory-
dependent. Nordlinger's paper is theory-dependent, but it
also introduces a fully worked-out proposal for dealing
with a complex syntactic phenomenon. The insights presented
in the remaining papers are embedded in the theoretical
framework to varying degrees, but all discuss data which
raise problems for any account of argument realization.
References:
Alsina, Alex (1997) A theory of complex predicates:
Evidence from causatives in Bantu and Romance. In Alex
Alsina, Joan Bresnan & Peter Sells (eds) Complex
Predicates, 203-246. Stanford CA: CSLI Publications
Bresnan, Joan W. (2001) Lexical-Functional Syntax. Oxford:
Blackwell Publishers
Carter, Richard (1976) Some linking regularities. In Beth
Levin & Carol Tenny (eds) On Linking: Papers by Richard
Carter. Cambridge MA: Center for Cognitive Science MIT
(Lexicon Project Working Papers no.25)
Halvorsen, Per-Kristian and Ronald M. Kaplan. 1988.
Projections and semantic description in Lexical-Functional
Grammar. In Proceedings of the International Conference on
Fifth Generation Computer Systems (FGCS-88), 1116-1122,
Tokyo, Japan. Reprinted in Mary Dalrymple, Ronald M. Kaplan,
John Maxwell, and Annie Zaenen, eds., Formal Issues in Lexical-
Functional Grammar, 279-292. Stanford: CSLI, 1995
Levin, Beth & Malka Rappaport Hovav (1995) Unaccusativity:
At the Syntax-Lexical Semantics Interface. Cambridge MA:
MIT Press
Matsumoto, Yo (1996) Complex Predicates in Japanese: A
Syntactic and Semantic Study of the Notion "Word". Stanford:
CSLI Publications
Nordlinger, Rachel (1988) Constructive Case: Evidence from
Australian Languages. Stanford: CSLI Publications
Ramchand, Gillian (1997) Aspect and Predication. Oxford:
Oxford University Press
Rouveret, Alain (1994) Le syntaxe du gallois. Paris:
Editions CNRS
Simpson, Jane (1983) Resultatives. In Lori Levin, Malka
Rappaport & Annie Zaenen (eds) Papers in Lexical-Functional
Grammar, 143-157. Bloomington IN: Indiana University
Linguistic Club
Simon Musgrave is a post-doctoral researcher at the
University of Leiden. His doctoral thesis is a study of
non-subject arguments in Indonesian, using LFG as the
theoretical framework. He is currently working on a cross-
linguistic database for the Spinoza Project, Lexicon &
Syntax, and is part of the East Indonesia research group
within the project.