LINGUIST List 16.2503

Mon Aug 29 2005

Review: Syntax/Typology: Carnie, Harley & Dooley (2005)

Editor for this issue: Megan Zdrojkowski <>


        1.    Asya Pereltsvaig, Verb First: On the syntax of verb-initial languages

Message 1: Verb First: On the syntax of verb-initial languages
Date: 27-Aug-2005
From: Asya Pereltsvaig <>
Subject: Verb First: On the syntax of verb-initial languages

EDITORS: Carnie, Andrew; Harley, Heidi; Dooley, Sheila Ann
TITLE: Verb First
SUBTITLE: On the syntax of verb-initial languages
SERIES: Linguistik Aktuell/Linguistics Today 73
PUBLISHER: John Benjamins
YEAR: 2005
Announced at

Asya Pereltsvaig, Department of Linguistics, Cornell University

This volume is a collection of papers from a workshop of verb-first
languages that took place at the University of Arizona, Tucson in February
2003. It offers a valuable contribution to our understanding of the syntax
of verb-first languages because it presents the most recent cross-
linguistic research on the subject from a variety of theoretical
perspectives. This book is of interest to both theoretical syntacticians
and typologists, as well as scholars who study the particular grammars of
verb-first languages, including Celtic, Zapotec, Mixtec, Polynesian,
Austronesian, Mayan, Salish, Australian, and Nilotic languages. It can
also be used as a text for advanced syntax courses (in fact, I will be
using several papers from this volume for my seminar on word order
derivation). The book opens with a short summary of the issues and the
following articles, written by the volume's editors: Andrew Carnie, Heidi
Harley, and Sheila Ann Dooley. It also contains a list of abbreviations, a
joint list of references, and an index; endnotes follow each separate


The main assumption shared by most of the contributors in this volume is
that verb-first orders (including Verb-Subject-Object, or VSO, and Verb-
Object-Subject, or VOS orders) are derived from a different underlying
order, namely Subject-Verb-Object (SVO). In addition to Kayne's (1994)
Universal Base Hypothesis, this assumption has been driven by evidence of
VP constituency in VSO languages. Yet, it should be noted that this
assumption is not itself uncontroversial; it has recently been challenged
by Borsley (in press); in fact, several contributions in this volume
reconsider this assumption and provide new evidence for it. Given this
assumption, however, the main issues considered by the contributors to
this volume are whether verb-first orders are derived by V-movement or VP-
movement and whether different verb-first languages should be analyzed in
the same fashion. Thus, the classical analysis of VSO languages, stemming
from the work of Emonds (1980), involves head-movement of V to some
functional projection higher than the subject (such as Tense or
Complementizer, depending on the analysis, or possibly on the language
considered). However, recent work on verb-first languages (e.g., Massam
2000, Rackowsky and Travis 2000) explored a different approach whereby
verb-first orders (both VSO and VOS) are derived by a movement of VP
remnant, that is a VP from which the subject (in VOS orders) or both the
subject and the object (in VSO orders) have been extracted
(or "evacuated"). If the remnant-VP approach is on the right track, it
means that VSO and VOS languages are more similar in their structure than
previously thought.

The article by Sandra Chung ("What fronts? On the VP-raising account of
verb-initial order", pp. 9-29) addresses both of these main issues: is the
verb-first order derived by VP-movement (as opposed to V-movement) and are
VSO and VOS languages subject to the same analysis? First, she considers
evidence for movement analysis of the VO string in VOS languages and
argues that while some languages (e.g., Seediq and Malagasy) conform to
the predictions of the VP-movement analysis, other languages (such as
Chamorro) are less amenable to such an analysis. Then, she turns to VSO
languages and raises an important question, largely ignored by previous
works that argued for VP-movement analysis (e.g., Massam 2000, Rackowsky
and Travis 2000), namely what motivation is there for extracting arguments
(and everything else apart from the verb itself) from the VP prior to
remnant VP movement? Although she does not give a definitive answer to
this question, she sets important goals for future research in this area.

Henry Davis in his "Coordination and constituency in St'at'imcets
(Lillooet Salish)" (pp. 31-64) takes up the issue of whether the
assumption that verb-first orders are derived from an underlying SVO order
is valid. He re-examines the evidence for subject-object asymmetries and
VP constituency in St'at'imcets and concludes that all tests except
coordination point to a hierarchical structure, with an underlying VP
constituent. As regards coordination, he concludes that its special status
in St'at'imcets is a property of coordination in general and not of
St'at'imcets in particular.

Yuko Otsuka's article "Two derivations of VSO: A comparative study of
Niuean and Tongan" (pp. 65-90) challenges the idea that all VSO
languages/orders are derived in the same fashion and claims that even such
closely related (Polynesian) languages as Niuean and Tongan require
different analyses: Niuean VSO order is derived by remnant VP-movement and
Tongan VSO order -- by head movement of V-to-T-to-C. The technical
implementation of this analysis is in saying that the two languages differ
in the nature of the Tense's EPP feature: in Niuean it is [Pred], while in
Tongan it is [D]. The important conclusion of this paper is that an
analysis that works for one VSO language cannot be immediately extended to
any and all VSO languages.

Felicia Lee ("Force first: Clause-fronting and clause typing in San Lucas
Quiaviní Zapotec", pp. 91-106) develops the line of analysis that relies
on string-vacuous movement of large constituents and argues that covert
clausal movement is both syntactically and semantically motivated in San
Lucas Quiaviní Zapotec. Her evidence comes from the existence of similar
overt movement in the language and semantic/interpretative constraints on
constructions that require that high projections in the left periphery be
filled. Thus, in the absence of base-generated particles in ForceP (such
as yes/no markers), remnant VP movement must apply, bringing the remnant
VP into the specifier of ForceP.

Thus, Lee's conclusions are at odds with those of the next article, "V1
and wh-questions: a typology" by Kenji Oda (pp. 107-133), who argues that
verb-first languages derived by VP-movement are banned from introducing
interrogatives by T-to-C movement. The only alternative available in this
languages, according to Oda, is the cleft strategy for fronting wh-items.
Oda further applies this conclusion to Irish and argues that Irish wh-
questions are pseudo-clefts; given this and the conclusions of the first
part of the article, it is concluded that Irish VSO order is best analyzed
with VP-movement rather than V-movement (but see Carnie and Harley,
forthcoming, for arguments against VP-movement analysis of Irish).

Dirk Bury in his article "Preverbal particles in verb-initial languages"
(pp. 135-154) seeks to explain the generalization that verb-initial
languages have preverbal particles. To do so, he proposes a model of
syntax where structural representations only state the dominance relations
between categories and the size of a given clause is determined by head
and specifier/adjunct positions that need to be merged. In his system, V-
movement is taken to be a PF (Phonological Form) operation which spells
out a verb either in the position of a derived head or in the position of
an independent category. While derived heads have no overt material and
always require a filled specifier, independent categories do contain
phonological material and in such structures the moved verb will appear
adjacent to some other overt material.

James McCloskey ("A note on predicates and heads in Irish clausal syntax",
pp. 155-174) turns his attention to the most-widely studied VSO language,
Irish. He considers data from ellipsis and coordination and argues that,
regardless of whether V-movement or VP-movement turns out to be the
correct analysis for Irish verbal clauses, clauses with non-verbal
predicates must be analyzed with head movement which raises at least
adjectival heads (and optionally prepositional heads) from the predicate
to a higher inflectional position. (Non-verbal predicates in VSO languages
are also considered briefly in Otsuka's article.)

Arthur Holmer ("Seediq: Antisymmetry and final particles in a Formosan VOS
language, pp. 175-201), like Davis, provides further support for the
validity of the assumption that verb-initial languages are derived from an
underlying SVO order, as proposed by Kayne (1994). His evidence comes the
occurrence of final particles in a VOS language, Seediq. He also
contributes to the discussion of whether verb-initial orders in different
languages are derived the same. Like Chung and Otsuka, Holmer argues that
the answer to this question is negative: Seediq and Tagalog differ in the
position of particle (final particles in the former and second position
particles in the latter) and hence differ as to the derivation of their
verb-initial orders.

Lisa deMena Travis in her article "VP-internal structure in a VOS
language" (pp. 203-224) considers how VP-ellipsis works in verb-first
languages. She shows that VP ellipsis in Malagasy (a VOS language) is
significantly different from VP-ellipsis in both an SVO language like
English and a VSO language like Irish. To account for these facts, Travis
argues that the basic mechanics of VP-ellipsis are the same across
languages, and the differences between Malagasy, on the one hand, and
English and Irish, on the other hand, arise from two considerations: (i)
the structure of the VP in Malagasy is created via iterative predicate
fronting and (ii) Malagasy VP-ellipsis involves an extra step of specifier-
head licensing that is missing in VSO and SVO languages.

Diane Massam's article "Lexical categories, lack of inflection, and
predicate-fronting in Niuean" (pp. 227-242) uses the previously developed
analysis of Niuean predicate fronting (including deriving its VSO order
via remnant VP-movement) to shed new light on the issue of lexical
categorization. She claims that Niuean verbs are not morphosyntactic verbs
in the same sense that English verbs are; instead, they are participial in
nature, have no features for finiteness and tense, and as such do not
establish a relation with INFL (Inflectional Head), undergoing predicate
fronting instead. The claim that Niuean does not have "proper" verbs is
especially interesting in light of Baker's (2003) recent claims that verb,
noun and adjective are universal lexical categories present in all

David Gil's article "Word order without syntactic categories: How Riau
Indonesian does it" (pp. 243-263) continues with the theme of lexical
categorization in verb-first languages. Gil bases his work on a previous
claim that Riau Indonesian has no distinction between lexical categories
such as verb, noun and adjective. Yet this language is known to show many
of the correlates of verb-first syntax, such as postnominal adjectives.
The question addressed in this article is how such statements can be made
without reference to lexical categories. The answer that Gil provides is
to state (most of the) word order observations in terms of a single
principle: heads precede their modifiers, with the residue of word-order
facts accounted for in terms of two additional principles: iconicity and
information flow. Thus he argues that the verb-first nature of Riau
Indonesian is epiphenomenal. It remains to be seen if the same claims can
be extended to other verb-first languages.

Mélanie Jouitteau ("Nominal properties of vPs in Breton: A hypothesis for
the typology of VSO languages", pp. 265-280) looks at the parallels
between CPs and DPs and argues that they are due to the fact that Breton
clauses (but not clauses in languages like English) have a [+D] feature on
the little v category. Using this idea, she provides an account for the
genitive case assignment system on internal arguments, the distribution of
preverbal prepositions and the fact that verbs appear to show Case-filter

Hilda Koopman ("On the parallelism of DPs and clauses: Evidence from
Kisongo Maasai", pp. 281-301) also turns her attention to the parallelism
between DPs and CPs; however, her claim is that the parallelism is due to
the fact that DPs in Kisongo Maasai are relative clauses with the [D CP]
structure. Furthermore, she shows that the differences between DPs and CPs
derive from independently motivated causes.

Loren Billings ("Ordering clitics and postverbal R-expressions in Tagalog:
a unified analysis?", pp. 303-339) takes up the issue of whether the
notion of subjecthood has any meaning (beyond a purely semantic one) in
Austronesian languages (specifically, Tagalog). His answer to this
question is positive: under his account, the possibility of both VSO and
VOS orders with Actor voice is due to the fact that in one of these orders
a proper name stands in what is normally a position reserved for pronouns.

Monica Macaulay's contribution ("The syntax of Chalcatongo Mixtec:
Preverbal and postverbal, pp. 341-366) is two-fold: first, she introduces
a new language into the linguistic discussion and describes the relevant
facts of Chalcatongo Mixtec; second, she closely examines the distribution
of the "left periphery elements" (which happen to be preverbal in this
otherwise verb-first language), including topic and focus. The difference
between topic and focus subjects in Chalcatongo Mixtec is that the former
is always doubled by a pronominal clitic, while the latter never is.
Macaulay's account for this is that the focused element is taken to be
moved to preverbal position, while the topic is taken to be base-generated

Mary Laughren, Robert Pensalfini and Tom Mylne ("Accounting for verb-
initial order in an Australian language", pp. 367-401), like Macaulay, Gil
and Billings, argue that syntax-external factors drive the clause-initial
placement of the verb. Specifically, they identify factors such as focus
and information structure as relevant in determining word order in Wanyi,
a verb-first language from Australia. (I might add here that focus and
information structure are also crucial in determining the word order in at
least some of the so-called "free word order languages", such as Slavic


This books presents valuable insights into the syntax of a variety of verb-
initial languages and as such is a great contribution to our understanding
of syntax. Accounting for word order across languages is one of the main
goals of syntactic theory, yet it is the word order problems that often
present the toughest challenges for syntacticians. The research
represented in this book goes a long way in elucidating the issues related
to a particular subset of word orders, those where the verb comes first.
As such, this book not only provides an overview of the cutting-edge
research on this subject, but also sets goals for future research.

One such goal and probably the biggest challenge in this domain of inquiry
today, to my mind, is buttressing the proposal that verb-initial orders
(be they VOS or VSO) are derived by remnant VP-movement; although this
analysis has already gained many proponents, it cannot be viewed as a
viable alternative to V-movement analyses as long as the question of
independent motivation for the creation of the remnant remains open.
Although this issue comes up in the first article in the volume (by Sandra
Chung), nowhere in the volume (or in fact anywhere else in the verb-first
literature) was I able to find a satisfactory answer to this question.
Since the remnant VP-movement itself is said to be triggered by the EPP,
the latter cannot be the motivation for the movement(s) that take
arguments outside the VP, thus creating the remnant. The fact that PP
arguments as well as DP arguments vacate the VP strongly suggests that
Case is not to blame either (see Chung's article in this volume, also
Rackowsky and Travis 2000). So what is it that forces everything apart
from the verb itself to vacate the VP in the first place? Without the
answer to this question, the remnant VP-movement analysis lacks in
explanatory adequacy and is nothing more than a mechanical account of what
moves where in order to get the word order right at the end of the


Baker, Mark C. (2003) Lexical Categories. Verbs, Nouns, and Adjectives.
Cambridge, UK: Cambridge University Press.

Borsley, Robert D. (in press) "On the nature of Welsh VSO clauses". Lingua.

Carnie, Andrew and Heidi Harley (forthcoming) "Clausal Architecture: the
Licensing of Major constituents in a verb initial language". Ms.,
University of Arizona.

Emonds, Joseph (1980) "Word Order and Generative Grammar". Journal of
Linguistic Research 1: 33-54.

Massam, Diane (2000) "VSO and VOS: Aspects of Niuean word order". In
Andrew Carnie and Eithne Guilfoyle (eds.) The Syntax of Verb Initial
Languages. Oxford: Oxford University Press. Pp. 97-116.

Rackowsky, Andrea and Lisa Travis (2000) "V-initial languages: X or XP
Movement and Adverbial Placement". In Andrew Carnie and Eithne Guilfoyle
(eds.) The Syntax of Verb Initial Languages. Oxford: Oxford University
Press. Pp. 117-143.


Asya Pereltsvaig is a Mellon Postdoctoral Fellow at the Department of
Linguistics of Cornell University. She is interested in derivations of
various word orders (both in clauses and in noun phrases). Most recently,
she has been working on the problems posed by word orders in noun phrases
in Slavic and Semitic languages.