EDITORS: Bowerman, Melissa; Brown, Penelope
TITLE: Crosslinguistic Perspectives on Argument Structure
SUBTITLE: Implications for Learnability
PUBLISHER: Routledge (Taylor and Francis)
Elena Tarasheva, Department of Linguistics, New Bulgarian University
The book is a collection of papers presented at a workshop on Argument Structure
held at the Max Planck Institute for Psycholinguistics in 1998. The researchers
draw evidence from a number of languages in support of either one of two
hypotheses: children learn their mother tongue by observing the linguistic
environment of verbs (syntactic bootstrapping, Gleitman 1990) or by fitting
words into semantic scenarios (semantic bootstrapping, Pinker 1984). The volume
can be of interest to specialists in first language acquisition, relational and
universal grammarians and to language typologists and people interested in
The articles have been grouped in three sections. The first section, entitled
''Verb Meanings and Verb Syntax: Crosslinguistic Puzzles for Language Learners'',
addresses the issue of whether verbs systematically appear in syntactic
environments which reveal their meaning and whether the semantic 'make-up' of
verbs conforms to the respective universal frames.
In her chapter, entitled ''A Person, a Place or a Thing? Whorfian Consequences of
Syntactic Bootstrapping in Mopan Maya'' (pp 29-49), Eve Danziger shows that in
Mopan Maya verb-like meanings, such as 'jump' and 'run', act as nouns, thus
obliterating the presumed universal distinction between nouns and verbs. In
spite of this fact, Danziger maps the structure of the language into learning
mechanisms which make it possible for children to learn the meanings through the
syntactic environments in which they occur. Thus the author endorses syntactic
bootstrapping at the expense of the noun/verb distinction which, in her opinion,
needs to be reconsidered.
In chapter 3, ''The Pitfalls of Getting from Here to There: Bootstrapping the
Syntax and Semantics of Motion Event Coding in Yukatek Maya'' (pp 49-69), Jurgen
Bohnemeyer discusses the fact that in Yukatek verbs expressing motion take
arguments which convey 'referential ground' - an argument type alien to motion
verbs in most European languages. He argues that children could not possibly
tune in to this way of framing motion unless they examine the morphosyntactic
properties of the verbs, which, in effect, discards semantic bootstrapping as
the explanation of language acquisition. However, the frequency of arguments
denoting 'referential ground' is not as high as expected in naturally occurring
speech and Bohnemeyer concludes that semantic and syntactic bootstrapping work
in tandem, facilitating and complementing each other in a dialectic relationship.
In the next chapter, ''Making Sense of Complex Verbs: On the Semantics and
Argument Structure of Closed Class Verbs and Coverbs in Jaminjung'' (pp 69-89),
Eva Schultze-Berndt reaches the same conclusion exploring co-verbs in Jaminjung,
an Australian Aboriginal language. Verbal meanings tend to be expressed by a set
of simple verbs which take a number of structures modifying the basic meanings.
Exploring the nature of these co-verbs, the researcher shows that children can
learn their meaning neither from the complex syntactic environments, nor from a
semantic frame. Syntax and semantics should function in tandem. Schultze-Berndt
is also led to believe that similar real-world scenes can not be described by
verbs with a common argument structure in different languages.
In chapter 5, entitled ''Figure-Ground Indeterminacy in Descriptions of Spatial
Relations: A Construction Grammar Account'' (pp 89-111), Sotaro Kita tackles the
problem that some languages allow saying both 'the skewer pierced the meat' and
'*the meat pierced the skewer', which feature he calls 'figure-ground
indeterminacy'. On the basis of Japanese, English and Likpe, he concludes that
Figure-Ground indeterminacy occurs with different verbs in different languages,
moreover pragmatic reasons can lead to various uses of the same verb. Therefore,
children need to learn to constrain the use of a verb on the basis of their
experience, without an inherent syntactic or semantic rule to guide their
choices. Kita admits that little research is available into this area of
language acquisition, and into this aspect of verb argument structure.
In her chapter, entitled ''Learning Verbs without Boots and Straps? The Problem
of 'Give' in Saliba'' (pp 111-141), Anna Margetts points to data from Saliba, a
Western Oceanic language, which she thinks are suitable for constructing a test
for the syntactic and semantic bootstrapping hypotheses. In Saliba, two verbs
express meanings associated with 'give': one is used for the first and second
person, the other for the third person. The latter satisfies the Universal
Grammar claim that such verbs take three arguments: a giver, a given object and
a recipient, but the former occurs only with an object given. The type of
acquisition mistakes can provide evidence whether children learn the verbs
proceeding from a semantic frame or from the arguments appearing with the verb.
So far, learning Saliba has not been sufficiently researched to yield data for
the test, but still Margett's material delivers a severe blow to the assumption
that a universal alignment between event types and syntactic structures exists.
The second section of the book is entitled ''Particpants Present and Absent:
Argument Ellipsis and Verb Learning'' and the chapters present research how
children learn verb meanings when in speech the arguments are often omitted in
the surface structure.
In chapter 7, ''Same Argument Structure, Different Meanings: Learning 'Put' and
'Look' in Arrernte'' (pp 141-167), David Wilkins tests the hypothesis that 'put'
and 'look' belong to different semantic classes and should thereby appear in
different syntactic frames. Material from Arrernte, an Australian Aboriginal
language, shows that both are in fact three-place verbs, contrary to the
syntactic bootstrapping hypothesis. The analysis of corpora leads Wilkins to
hypothesize that children do not automatically take the appearance of arguments
as an indication that the verbs belong to a specific semantic class with its
respective argument structure. Alignments between a syntactic frame and a
semantic meaning can be more or less natural and children tend to base their
decisions on more natural alignments rather than on less natural ones.
Chapter 8, ''Verb Specificity and Argument Realization in Tzeltal Child Language''
(pp167 -191), presents Penelope Brown's study of semantically complex verbs in
Tzeltal, a Mayan language. The data challenge the view that nouns are easier to
learn than verbs in first language acquisition. Because children tend to learn
Tzeltal verbs first, as carriers of a greater amount of semantic information,
Brown endorses what is known as language-typology bootstrapping (Slobin 2001):
the organization of a language determines which structures are acquired first.
However, in her opinion, the specialization occurs at later stages of language
learning: initially, words are learned as syntactic frames around a semantic
core. Brown also observes that the direct objects of verbs incorporating the
meaning of the respective objects ('eat crunchy food') are realized less
frequently by both adults and children than the direct objects of verbs with
more general meanings ('eat no matter what'). Therefore, patterns of argument
ellipsis are related to the semantics of a verb by a link that is possibly
inherent for the learners.
In the final chapter of the second part, ''Interacting Pragmatic Influences on
Children's Argument Realization'' (pp 191-213), Shanley Allen addresses the issue
of which discourse-pragmatic factors are taken into consideration by children
when they learn to drop verb arguments. Based on corpora from Inuktitut and
Korean, the researcher explores factors such as newness, contrast, absence and
person. Allen collates her data with other investigations and concludes that
children are most sensitive to the informativeness factors. Also, factors
function cumulatively to impact overt argument realization.
The third part of the book, ''Transitivity, Intransitivity, and their Associated
Meanings: A Complex Work-Space for Learnability'', contains diverse articles on
transitivity and intransitivity.
If Perlmutter's Unaccusativity Hypothesis (1978) is correct, intransitive verbs
should belong to one of two classes: unergatives and unaccusatives and children
gradually learn to expect either. However, James Essegbey demonstrates in
chapter 10, ''Intransitive Verbs in Ewe and the Unaccusativity Hypothesis'' (pp
213-231), that in Ewe, a West African language, the intransitive verbs in fact
belong to a single class marked by a common property, namely their inability to
express a causal meaning. This situation poses the question of how children
determine the argument structure of the verbs. Syntactic bootstrapping may help,
according to Essegbey, but following the acquisition of the specifics of control
In chapter 11, ''He Died Old Dying to be Dead Right: Transitivity and Semantic
Shifts of 'Die' in Ewe in Crosslinguistic Perspective'' (pp 231-255), Felix K.
Ameka discusses material from Ewe related to the verb glossed as 'die'. He
supposes that the learning of the Ewe verb would digress from that of 'die'
verbs in other languages because of its specific semantic and syntactic
properties: the Ewe verb can be used with two surface nouns and often undergoes
semantic shifts imparting meanings such as 'remove the function'. It is the case
that meanings associated with the two-place use are identical with the meanings
of the one-place use and conversely, the two-place uses differ significantly
among themselves. Ameka establishes that the decisive factor distinguishing one
set of meanings from another is the feature 'animate'. Therefore, syntactic
bootstrapping only works when children have acquired the specifics of the verb's
arguments, especially animacy.
In chapter 12, ''Acquiring Telicity Crosslinguistically: On the Acquisition of
Telicity Entailments Associated with Transitivity'' (pp 255-279), Angeliek van
Hout reviews several experiments where children acquire the telicity of verbs,
i.e. their specific property to denote a culmination of the verbal activity. In
Slavic languages telicity is encoded in the predicate and acquired with greater
ease and at a younger age, unlike the Germanic languages in the study -
English, Dutch, and Finnish - where telicity is computed on the basis of the
joint properties of the verb and object. 'Compositional telicity', as van Hout
calls this type, poses more problems for children. Bybee's (1985) relevance
principle explains why telicity is more relevant to verbal categories, which
makes it easier in its predicate realizations.
Melissa Bowerman and William Croft, in their chapter entitled ''The Acquisiton of
the English Causative Alternation'' (pp 279-309), study data from a longitudinal
corpus of children's speech to establish how causative meanings are derived from
a number a verbs in English. According to the nativist proposal (Baker 1979),
inborn constraints prevent children from presuming that 'she died it' means 'she
killed it'. Alternatively, Bowerman and Croft suggest that children discover
which verbs allow causative transformation in several steps: firstly, they learn
individual verbs; then they generalize across a wide range of forms, including
causative ones; later, errors abate under the strength of evidence that
high-frequency causatives exist; finally, semantic subclasses develop, which
limits causative conversions to the permissible ones only. The actual material,
however, offers little justification for either model, therefore the hypothesis
of entrenchment remains the only plausible explanation, i.e. children learn
which verbs allow causative use by repeatedly hearing them in intransitive
frames until the association between verb and frame prevails in the child's own
Chapter 14, ''What Adverbs Have to do with Learning the Meaning of Verbs (pp 309-
331) is by Angelika Wittek. She investigates the acquisition of transitive verbs
expressing a change of state in English and German. A series of experiments
shows that for German-speaking children the end state of the verbal action does
not present a critical meaning component. For her respondents this group of
verbs is semantically similar to verbs denoting processes, which causes a
problem for both the syntactic and semantic bootstrapping proposals. The use of
an adverbial, however, serves as sufficient disambiguation. Therefore, Wittek
proposes her own Adverbial Modification Cue hypothesis. According to this
hypothesis, the use of adverbs meaning 'again' is a cue from which children
learn that certain verbs entail an end state.
Finally, in chapter 15, ''Event Realization in Tamil'' (pp 331- 357), Eric
Pedersen also studies end states but with reference to Tamil. In this language
the semantics of the verbs allows speakers to say 'he broke the coconut but it
did not break'. The strength of the entailments for verbs of this semantic type
varies across languages and uses. Experiments lead Pedersen to believe that
there may be some universal tendency to use change-of-state verbs for unrealized
events, but the processes are specific for each language. An overarching
systematicity to the adult patterns reveals a changing sensitivity in child
usage in their particular language.
The main merit of the book is that it puts together research by specialists in
different areas aimed at solving the problem of how children learn to speak
their mother tongue. Studies by linguists, neuroscientists, language acquisition
specialists and psychologists can be expected to give a wide range of evidence
in favor of or against the proposed hypotheses. Moreover, most of the authors
are established names in their areas and continue their efforts from previous
Secondly, the book overcomes the all-too-familiar Eurocentric perspective on
languages by including African, Australian, and Asian languages in the main bulk
of the presentations. It may be the case that those languages are spoken and
studied by fewer people, but if claims are to be made for a Universal Grammar,
then a wider range of languages needs to be considered.
A praiseworthy feature of the book - probably due to its cross-disciplinary
nature - is the accessible style of the articles. With such narrowly
specialized disciplines, one would have expected a concentration of terminology
impenetrable to the non-specialist. However, all the research is presented in a
reader-friendly way, without too much terminological complication and
unfathomable references to ''must-be-familiar'' facts and quotations. Especially
worth reading is the introduction to the book (pp. 1-29), where Melissa Bowerman
and Penelope Brown introduce the subject of bootstrapping in plain and concise
The chapters introduce a variety of research methods - from typological
outlines, deep structure analyses to statistical performance tests. It feels,
however, that when convincing results are needed, researchers are turning more
and more often to corpora. Actual speech production gives ample material for
speculation, appreciation, classifications and re-classifications, as can be
seen from this volume.
As for building an argument that has the power to persuade, the book gives food
for thought. On the whole, none of the language descriptions in the collection
gives evidence which would formally endorse semantic or syntactic bootstrapping.
At best, the authors conclude that syntactic and semantic bootstrapping work in
tandem (because neither seems possible), or both are superceded by
language-typology bootstrapping, the relevance principle, or entrenchment. Even
worse, bootstrapping is shown to work provided that the mainstay noun/verb
distinction is violated. Having read the book, an unprejudiced reader has
reasons to presume that the plausible explanation for first language acquisition
seems to be the entrenchment hypothesis (Braine 1971), because all the studies
show that in the end children learn from repeated exposure to language
phenomena. Moreover, the material strongly disproves a match between syntactic
frames and semantic meanings across languages - this theme recurs in all the
papers. Paradoxes of this kind call attention to the basic question to all
research - what actually serves as proof for a theory and what can topple it?
Nevertheless, the book makes a wonderful read, its educational value is immense
and the labor invested in it has been worth every minute. Bootstrapping helps
understand various language phenomena, even when disproved.
Baker, C.L. 1979. Syntactic Theory and the Projection Problem. _Linguistic
Inquiry_, 10 (44): 533-581.
Braine, M.D.S. 1971. On Two Types of Models of the Internalization of Grammars.
In D.I. Slobin (ed.) _The Ontogenesis of Grammar_, pp 153-186. New York:
Bybee, J. 1985. _Morphology: A Study of the Relation between Meaning and Form_.
Amsterdam: John Benjamins.
Gleitman, L. 1990. The Structural Sources of Verb Meanings. _Language
Acquisition: A Journal of Developmental Linguistics_, 1 (1): 3-55.
Perlmutter, D.M. 1978. Impersonal Passives and the Unaccusative Hypothesis. In
_Proceedings of the Fourth Annual Meeting of the Berkeley Linguistics Society_,
157-189. Berkeley, C.A.: Berkeley Linguistic Society, University of California.
Pinker, S. 1984. _Language Learnability and Language Development: The
Acquisition of Argument Structure_. Cambridge, MA: Haarvard University Press.
Slobin, D. I. 2001. Form-function relations: How do children find out what they
are. In M. Bowerman, and S.C. Levinson (eds) _Language Acquisition and
Conceptual Development_, 406-449. Cambridge: Cambridge University Press.
ABOUT THE REVIEWER
Elena Tarasheva has publications discussing corpus linguistics and developing
intercultural competence. She has been teaching English as a foreign language
for more than 20 years to all sorts of audiences. Currently she is assistant
professor at the New Bulgarian University in Sofia, teaching courses in
Computational linguistics, English Phonetics and Phonology and Media English.