AUTHOR: Gillian Catriona Ramchand
TITLE: Verb Meaning and the Lexicon
SUBTITLE: A First-Phase Syntax
SERIES TITLE: Cambridge Studies in Linguistics 116
PUBLISHER: Cambridge University Press
YEAR: 2008
Stef Spronck, Department of Linguistics, Research School of Pacific and Asian
Studies
SUMMARY
Argument structure and the division of labor between the grammar and the lexicon
are issues every theory of grammar will have to address but both problems have
gained particular significance with the introduction of constructionist
approaches to language, particularly the influential Goldberg (1995). Any study
that aims to contribute to this key area of theoretical linguistics steps into
the tradition of a vast literature encompassing verb classification, semantic
roles, core arguments, external arguments and non-arguments, the
characterization of lexical knowledge and the delineation of structure and
meaning. In the volume under review, Ramchand treats the lexical/syntactic
representation of verbs within a broad perspective that addresses central claims
and sentence types that have been put forward in studies on valency and aspect
from a range of theoretical approaches and places her analyses in a Minimalist
syntactic framework she refers to as a first-phase syntax.
The volume is structured as follows: in the introductory chapter Ramchand treats
a number of approaches to the lexicon with special reference to the status of
semantics, syntax and encyclopedic knowledge and classifies her proposal as
generativist-constructivist. The second chapter is devoted to a wide range of
sentence types Ramchand introduces as basic verb-argument structures, which in
chapter 3 she places in a syntactic model, a first-phase syntax. Chapter 4
applies this first-phase syntax to English verbs and demonstrates how they may
be classified on the basis of the position they occupy in the projected
syntactic structure. Chapters 5 and 6 take a more in-depth and cross-linguistic
look at two phenomena that receive specific treatment in the model,
results/paths and causitivization, respectively, and the programmatic chapter 7
explores further possibilities of the proposed analysis and indicates open
questions. I summarize each of these chapters in some detail below, comment
briefly on the terms 'constructionist', 'constructional' and 'constructivist'
and conclude with a general evaluation.
Chapter 1 Introduction
Ramchand sets the stage by putting forth the claim that ''[...] to the extent
that lexical behaviour is systematic and generalizable, this is due to syntactic
modes of combination [...]'' (1). Before substantiating this claim her first task
is to define the relation between the mental lexicon and the syntax, to which
she devotes the main part of this introductory chapter. In brief, Ramchand
distinguishes two conceptions of the nature of the lexicon, the
'lexical-thematic approach' and 'the generativist-constructivist approach' and
both have an extreme and a more moderate interpretation. Under the extreme
reading the lexical-thematic approach conceives of the lexicon as a bundle of
information in which each verb is stored with fixed information about its
arguments (e.g. valency, semantic roles, cf. most notably Dowty, 1990) which are
instantiated in the syntax through linking rules with ''no lexicon-internal
manipulations prior to insertion'' (8) (the 'static lexicon' view). A moderate
reading of the lexical-thematic approach leads to a view in which, although
verbs in the lexicon are thought of as carrying information about argument
structure these are determined by regular processes in the lexicon (the 'dynamic
lexicon'), e.g. when it comes to aspect specification. The
generativist-constructivist approach has two versions as well: one in which the
verb is stored without any specification of its argument structure and is just a
bundle of ''[...] cognitive and encyclopaedic information'' (9) (the 'naked roots
view'), and one in which at least some information about valency and semantic
roles is stored with the lexical entry (the 'well-dressed roots view'). It needs
to be observed that the main distinction between the static lexicon view and the
well-dressed roots view lies in the theoretical framework in which it is
embedded, decompositional versus constructivist (11). Ramchand does not readily
subscribe to any of these views. In her interpretation of the lexicon a lexical
entry is nothing more than ''[...] the memorized link between chunks of
conceptual structure and conditions of insertion'' (14). This cold be regarded as
a moderate version of the generative-constructivist view, in which verbs with
limited syntactic labels are associated with particular constellations in a
generative syntactic structure but it should be noted that these should not be
seen as properties of individual lexemes since similar patterns are found
cross-linguistically (13).
Chapter 2 The empirical ground
Although some verbs, e.g. 'eat' may be applied in a wide range of sentence types
(transitive, intransitive, causative, way-constructions etc.) the use of other
verbs is much more constrained. Where lies the source of these restrictions?
''The strategy I will pursue is first of all to reject the existence of formal
selectional features in the lexicon, but attempt to account for what rigidity
there is in terms of purely syntactic or categorial features, made possible by a
more articulated view of the functional sequence within the verb phrase [...]
The first step is to establish and motivate the primitives that are empirically
necessary in a decomposition of verbal meaning'' (22). In doing so, Ramchand
introduces a range of sentence types that have been adduced in the argument
structure literature on the basis of which she formulates the following
reoccurring semantic roles: initiator (''an entity whose properties/behaviour are
responsible for the eventuality coming into existence'') (24), undergoer
(''argument that is interpreted as undergoing the change asserted by [a] dynamic
verb'') (27), resultee (''direct argument related to [a] result state'' (33)), path
and rheme. The latter two are the most distinctive in Ramchand's proposal.
''[D]ynamic verbs have a part-whole structure, as defined as our human perception
of the notion of change'' (26), and these parts are 'paths'. Rhemes are ''objects
of stative verbs'' (34). Ramchand assumes that these are the basic semantic roles
of arguments, accounting for such verb features as 'causality' and 'telicity',
which she considers to be secondary.
Chapter 3 A first-phase syntax
This chapter gives a first overview of Ramchand's model, which she stresses to
be programmatic rather than definitive. On the basis of the sentence types
introduced in the second chapter she distinguishes three subevents in the
event-structure: ''a causing subevent, a process-denotiong subevent and a
subevent corresponding to result state'' (39). These are interpreted as three
projections: the initiation phrase (initP), of which the subject position is
taken by the subject of a cause, or INITIATOR. This phrase may head a process
phrase (procP), the subject of which is the subject of a process, or UNDERGOER,
which in turn may head a result phrase (resP), the subject of which is a
RESULTEE. The object position of the lowest phrase may both be taken by a PATH
or a RHEME. This structure is the first pillar on which the theory rests. Two
other main components are the notions 'merge' and 'monotonicity'.
In Ramchand's proposal the capitalized roles above are not monolithic, but can
be combined to form complex roles. Ramchand characterizes the syntactic
operation with which this is done through the central Minimalist term merge,
which she conciliates with more traditional interpretations of the theoretical
construct by stating that: ''if merge is conceived of as set information, then
nothing prevents a particular item from being a member of more than one set''
(59). (1-6) are examples of these 'pure' and 'combined' roles:
(1) The key opened the door ('the key' = initiator)
(2) The ball rolled (= undergoer)
(3) Ariel ate the mango ('the mango' = path)
(4) Katherine ran her shoes ragged ('her shoes' = result)
(5) The diamond sparkled (= undergoer-initiator)
(6) Katherine broke the stick ('the stick' = resultee-undergoer) (52-53)
Ramchand assumes a regular connection between the subevents in macro-events
(which can be 'decomposed' into the subevents indicated by the phase types) and
the semantic roles of the core and external arguments (which are derived from
argument positions in the structure). Based on formalizations proposed by Roger
Schwarzschild, Ramchand conceives of the relation between these two as property
sets that have to be 'monotonic' and ''a relation between two structured domains
is said to be monotonic if it preserves the ordering from one domain to the
other'' (49).
The remaining part of the study serves to show that this structure may indeed
account for the recurring structures and meanings that are found in verbal
predication.
Chapter 4 Deriving verb classes
The first-phase syntax relates in systematic ways to particular verb types,
although the association between the two should more be seen a pattern rather
than a one-to-one correspondence. The present chapter explores these patterns
and analyses a wide range of English sentence types that have been adduced in
the argument structure, construcionist and Aktionsart literature. Systematic
correlation between aspect/Aktionsart and certain constellations in the
syntactic structure are indicated, e.g. ''when a single lexical item identifies
both 'proc[ess]' and 'res[ult]' [...] the event expressed is punctual'' (77).
Chapter 5 Paths and results
In this chapter path and result sentences (cf. (3-4) above) are further
illustrated and path phrases are decomposed in a path phrase heading a place
phrase (following Ray Jackendoff's distinction between path and place
prepositions). Ramchand demonstrates the cross-linguistic relevance of the
proposed structure by further introducing constructions from Korean (in which
path phrases need not be introduced by a preposition), Swedish, Norwegian,
Russian (demonstrating the interaction between aspect prefixes and argument
structure in the language) and Indic languages.
Chapter 6 Causativization
The penultimate chapter of the volume picks out one particular construction in
Hindi/Urdu that has received some attention in the literature, the causative
construction, and Ramchand demonstrates that the first-phase syntax may account
for the alternations found in the language.
Chapter 7 Conclusion
After summarizing the study and further specifying the relation between tense,
aspect and verbal decomposition, the volume concludes with a number of open
questions, among which the place of case in the system and stative verbs.
On the terms 'constructivist', 'constructionalist' and 'constructionist'
The influence of construction grammar (esp. Goldberg, 1995) can be felt
throughout this study. ''I am attempting to implement an old idea in the light of
current, accumulated knowledge concerning the nature of 'lexical'
generalizations and patterns''(1). This ''current knowledge'' refers to the
constructionist approach, but it is the ''old idea'' which makes Ramchand's volume
a valuable and innovative contribution to the constructionist debate, although
the author explicitly distances herself from constructionism: ''the view proposed
here will be generative-constructivist in spirit, but not 'constructionist'''
(11). This reaction appears to be addressed to what Ramchand calls the 'radical
constructionalist approach' to the lexicon in which ''no lexical information is
present at all, but lexical items are inserted into syntactic contexts according
to compatibility with encyclopaedic and real-world knowledge'' (21). One may
question if this is an accurate depiction of Radical Construction Grammar
approaches in the sense of Croft (2001) in which the notion of insertion is
irrelevant since there is simply no fundamental distinction between
conventionalized meaning on the level of lexemes or on the level of phrases,
which reduces the difference of opinion to a rather more fundamental theoretical
position on the nature of the required operations in the grammar. But the main
distinction between her model and construction grammar models Ramchand aims to
stress is that ''unlike the 'constructional' grammar of Goldberg (1995)
[event-compositional] semantics will not be associated with arbitrarily large
syntactic objects, but constructed systematically on the basis of primitive
recursive syntactic relationships'' (16). It seems that Ramchand defines
'constructivist' as a systematic way of deriving meaning from structure, whereas
by 'constructional' she refers to surface structures which receive their meaning
from syntax internal operations, which, she argues, are overlooked in current
construction grammar approaches.
It may lead to some surprise to find that elsewhere the author does use the term
'constructional' to refer to her own model, cf. ''[b]ecause this is a
constructional [sic] system, the wide variety of different verb types and role
types will be derived from the different combinatoric possibilities of the
syntax'' (63). Here, the term refers to the entire syntactic structure from which
meaning is derived: ''the syntax with a basic templatic semantics is built up
autonomously, as one tier or dimension of meaning (a constructionalist view),
with the association to lexical content providing the other tier or dimension of
meaning'' (58). These remarks are not as inconsistent as Ramchand leads us to
believe in her first chapter, in fact, the notions are quite compatible: the
model outlined in the present volume is a specific (generative-constructivist,
if one prefers) account of constructional meaning and as such forms part of a
larger constructionist debate. I interpret the distinctions between the
different terms to be ideological rather than fundamental, referring to
particular stances in the debate, and will continue to use 'constructionist' to
denote any theoretical position that assumes a semantic dimension above the
level of lexemes whether this view is embedded in a cognitive,
structural-functionalist or generative approach to language structure.
EVALUATION
I will have to leave the judgment about the extent to which the volume
constitutes a contribution to the Minimalist Program to others, but the ways in
which in the proposed first-phase syntax multiple roles may be combined in a
Merge operation to form complex semantic constellations certainly seems an
interesting alternative to monistic Principles & Parameters proposals and fully
in the spirit of Minimalism as I understand it. But above all, Ramchand's study
opens a dialogue between constructionist approaches to grammar and formal
grammar models. If Minimalism is to be interpreted as a step towards convergence
in linguistic theory (assigning a central role to economy as many functionalist
approaches traditionally do and discarding transformations), just as
constructionist approaches to language structure might be seen, having come to
permeate formal and cognitivist-functionalist grammars alike, the volume is of
potential interest to a wide range of readers. It may serve as an excellent
starting point for opening up a discussion between currents in theoretical
linguistics about a topic every linguistic theory has to address.
Anyone who would be willing to engage in this debate will find in this volume a
fascinating study in which the author builds up her argument carefully,
unveiling her proposal layer by layer while raising profound and intriguing
questions in the process. The model she presents is descriptively elegant and
Ramchand applies it to an array of phenomena and structures, many of which have
not been previously treated in an integrated account.
Different aspects of the study may appeal to different readers depending on
their backgrounds. A functionalist reader may question why the phrases denoting
subevents and argument roles would have to be interpreted as syntactic positions
since basically they correlate with the semantic properties of argument-verb
constructions that have been adduced in the literature all along. Even they,
however would have to appreciate the clear predictions that can be made and
tested on the basis of the proposed structure. A more typologically oriented
reader may be alarmed by the classical way in which the presented analysis is
claimed to be universal: the author argues for the model based on English and
then applies it to a number of selected examples from other (mainly
Indo-European) languages. On the other hand, the volume does shed new light on
the non-English examples as well. For example, the proposed link between
argument structure and Russian aspect prefixes is an interesting and innovative
proposal in the study of aspect, although I would be curious to see how aspect
(imperfectivizing) suffixes in Russian would have to be treated under this view.
Finally, a cognitivist reader may object to interpreting notions like rheme and
path as purely linguistic concepts, which leads to such analyses as e.g. the
decomposition of a phrase like 'into Determiner Phrase' in a node 'place' with
the preposition 'in' and a node 'path' with the preposition 'to' (114). I would
hope, however, that readers from all these different parts of the field will
find their way to this stimulating study and will address the important
questions it raises.
REFERENCES
Croft, William. (2001) _Radical Construction Grammar, Syntactic Theory in
Typological Perspective_. Oxford: Oxford University Press.
Dowty, David. (1990) Thematic Proto-Roles and Argument Selection. _Language_,
Vol. 67, No. 3, pp. 547-619.
Goldberg, Adele E. (1995) _Constructions, A Construction Grammar Approach to
Argument Structure_. Chicago: The University of Chicago Press.
ABOUT THE REVIEWER
Stef Spronck is a PhD student at the Australian National University working on
different aspects of the grammar of the North-Western Australian language
Ngarinyin. His main research interests involve constructions of reported speech,
the function of grammatical categories in situated language and linguistic
theory, particularly constructionist and functionalist approaches to language
structure. His current research focuses on verbal morphology and verb
classification.
|
