AUTHOR: João Costa
TITLE: Subject Positions and Interfaces
SUBTITLE: The Case of European Portuguese
SERIES: Studies in Generative Grammar
PUBLISHER: Mouton de Gruyter
Michael Barrie, Department of Linguistics, University of Toronto
Summary: This manuscript is an updated version of the author's 1998
dissertation. Its main goals are to account for the variations in surface
word order and subject positions in European Portuguese (henceforth EP)
within a Minimalist framework (Chomsky, 1995). Of the six logically
possible combinations among subject, verb, and object only SOV is not
attested with the other five being found in specific discourse contexts,
which Costa describes. The implication of Costa's work is that syntax
operates independently of semantic or discourse factors. Specifically, he
proposes that syntax may make more than one possible construction available
to the interface; however, only one of these syntactically available
constructions will be chosen as an output form based on semantic and
discourse factors. I present a brief summary of each chapter, followed by a
critical evaluation of the work.
Chapter 1: The introductory chapter sets the scene for this monograph,
detailing the model of grammar that Costa adopts, namely Minimalism
(Chomsky, 1995) combined with Reinhart's (1995) proposals on the interface
(subsequently published as Reinhart, 2006) and Distributed Morphology
(Halle and Marantz, 1993). Costa also lays out some basic assumptions
regarding verb movement and the position of adverbs. Based on earlier work
of his, he assumes that the verb raises to the head of TP, but not as high
as the head of AgrP. He also assumes that monosyllabic adverbs left-adjoin
to VP, while other adverbs freely left-adjoin to any projection (within
certain limits, of course). Costa also sets out the main conclusion of his
study on word order and subjects in EP, namely that the subject can occupy
either SpecVP, SpecTP, or SpecAgrP as long as the following two conditions
are met: i) the position is made available by the syntax, ii) the position
does not violate any interface condition. Based on this fact about the
availability of subject positions, verb raising and discourse factors,
Costa derives the five possible word orders of EP (SVO, VSO, VOS, OSV, OVS)
and accounts for the lack of SOV.
Chapter 2: This chapter deals with the location of preverbal subjects in
EP. Costa's first order of business is to deal with popular view that overt
preverbal subjects in Romance languages are left-dislocated. Costa employs
a battery of diagnostics showing that preverbal subjects are indeed not
left-dislocated but rather occupy a canonical argument position. Briefly,
these include multiple preposing (left-dislocated elements are not rigidly
ordered, however subjects are with respect to other left-dislocated
elements), lack of reconstruction effects (unexpected if the subject is
left-dislocated to an A-bar position, but expected if the subject is in an
A-position), pronominal doubling, and the lack of preverbal subjects in
I-to-C contexts, among several other tests. Costa concludes that preverbal
subjects are in an A-position.
Chapter 3: This chapter deals with the location of postverbal subjects.
Specifically, Costa looks at VSO and VOS word orders, leaving OVS for later
discussion. Costa argues that VSO word order is the result of V-to-I
movement while the subject remains in situ. To do this, he argues against
the idea that VSO order arises by I-to-C movement across the raised
subject. To this end, he argues that I-to-C movement cannot be the source
for VSO order since VSO is found in embedded contexts with an overt
complementizer. He also argues that the subject in VSO remains in situ
based on adverb placement. (Recall above that Costa concluded that
monosyllabic adverbs appear only at the left edge of VP.)
Costa then turns his attention to VOS word order, claiming it is derived
the same as VSO order, except with object scrambling across the subject.
Costa first rejects the traditional analysis of VOS in which subjects are
right-dislocated on the basis of various syntactic diagnostics for
right-dislocated elements and further concludes that subjects in VOS are in
SpecVP on the basis of tests similar to those above. He then takes a
detailed look at scrambling in Dutch and German, drawing on several
similarities between scrambled objects in these languages and objects in
VOS order in EP in support of his claim the objects in VOS constructions
are scrambled. Costa also devotes a large portion of this chapter arguing
against a remnant movement analysis of VOS word order.
Chapter 4: This chapter deals with the interplay between focus and word
order in EP. Costa argues that certain focused objects are licensed in situ
rather than in the specifier of a Focus Phrase in the left periphery. Costa
contends that both kinds of focus are available in EP, but turns his
attention to in situ focus. It is in this chapter that Costa relates the
variety or word orders (specifically SVO, VSO, and VOS) to topic and focus
structures. Costa is concerned here with presentational focus, or that part
of the sentence that answers a wh-question. The author adopts the view that
syntactic operations can be triggered by prosodic factors and old
information must either be topicalized or move to a non-focus position as
the right edge of the clause (where nuclear stress falls) is a position of
focus. These facts taken together allow Costa to derive the various surface
word order found in EP. What's crucial here is that, following Costa's
arguments, the syntax makes various word orders possible; however, semantic
and pragmatic factors choose only one of these convergent derivations as
the final surface form. Finally, Costa briefly discusses phases following
more recent work in Minimalism (Chomsky, 2001) and copular constructions,
leaving various questions for future research. For instance, Costa comes to
the unconventional conclusion that vP is either absent or is a weak phase
in EP, but only in VSO contexts.
Chapter 5: One of the author's main goals is to eliminate optionality in
grammar. To this end, Costa deals with apparent cases of optionality in
unaccusatives and answers to multiple wh-questions. Unaccusatives are
challenging as they allow either SV or VS order in neutral discourse
contexts. Costa probes further into unaccusatives, noting that agreement is
optional with postverbal subjects, and suggests that VS word order encodes
identificational focus on the subject. In doing so, Costa offers some
interesting data involving overt expletives in colloquial EP. Costa then
goes on to discuss apparent optionality in answers to multiple
wh-questions, where both VSO and SVO orders are observed. Here, Costa
employs various diagnostics suggesting that VSO entails exhaustivity in the
response to multiple wh-questions while SVO does not entail any such
Chapter 6: This chapter deals with the interface between syntax and
morphology and how it relates to subject position. Costa adopts Bobaljik's
(1995) Distributed Morphology approach to the morphosyntax of verbal
inflection. Specifically, Costa adopts the view that morphological merger
between two heads is blocked by an intervening lexical material (except
adverbs). Recall from above that Costa assumes that the verb in EP raises
only as far as the head of TP. Since the verbal morphology appears in the
head of AgrP, the subject cannot appear in SpecTP or morphological merger
between the verb and the agreement would be blocked. Costa then argues that
if the verb raises to the head of CP (via AgrS) then SpecTP becomes
available for the subject.
Chapter 7: This chapter develops Costa's view of the interaction of syntax
and the interfaces. As mentioned above, Costa adopts the view that syntax
operates independently of discourse factors and pragmatics, and that the
syntax may create more than one convergent derivation from which the
interfaces choose the one that best suits its needs. Furthermore, the
author proposes that operations at the interface take place only as a last
resort if the syntax does not generate a derivation that it finds suitable.
Costa discusses data on English and EP ditransitives to support this view.
He argues, for example, that sentential stress may shift only as a last
resort if the syntax does not generate a derivation in which the focused
argument appears in the position of nuclear stress.
Chapter 8 presents a brief summary of the monograph and its conclusions.
Costa also includes an extensive appendix entitled ''On the nature of
agreement in European Portuguese.'' The author contrast EP with two
varieties of Brazilian Portuguese (BP1 and BP2) with respect to number
agreement, both within the DP and between the subject and verb. Briefly,
EP, BP1 and BP2 exhibit varying degrees of obligatory number agreement
(although person and gender agreement is generally obligatory in all
cases). Referring to earlier work in which it is claimed that the verb
raises to the head of TP in all three varieties under consideration here,
Costa pursues an analysis in which the [plural] morpheme is inserted
post-syntactically in the morphological component. (Recall that Costa
assumes a DM framework.) Technically, the [plural] morpheme is dissociated
in EP (in the sense of Embick and Noyer, 2001) but is a singleton in BP.
This distinction, along with a few others, derives the surface differences
in these three varieties without appealing to differences in verb raising.
The title of this monograph suggests that subject positions and the
interfaces with the syntactic component constitute the central thesis. An
equally important contribution of this volume, however, is Costa's thorough
description of the word order possibilities in EP and the discourse
contexts under which the various orders are found. Theoretically, Costa
advances a view of language in which syntax, morphology and the interfaces
are independent and do not affect one another. An important consequence of
this view for the author is that it is possible for the syntax to generate
multiple convergent derivations among which the interfaces select the one
that best fits the discourse context. The strength of Costa's argumentation
lies in the large number of diagnostics he employs to support his claims on
the clausal structure of EP and the positions of subjects in that language.
An important aspect of this study is its interest not only to scholars of
Romance syntax, but to a wider audience as well. For instance, Costa's
analysis has implications for the ongoing debate on the status of
morphology -- that is whether morphology is a completely autonomous domain
(Aronoff, 1994), whether it separate, but interacts with syntax (Halle and
Marantz, 1993, et seq.), whether it proceeds in parallel with syntax (Di
Sciullo, 2005), or whether morphology is does not exist at all, and what we
think of as morphology is done in the syntax (Julien, 2002, Starke, 2003).
While Costa presents a thorough empirical discussion of EP, augmented by
data from varieties of BP, which is supported by an impressive range of
diagnostic evidence there remain various puzzles and unanswered questions.
For instance, Costa assumes that monosyllabic adverbs adjoin to VP, while
other adverbs may optionally adjoin to virtually any XP. While the
empirical evidence given suggests this is a possible analysis, it is a
rather strange requirement of the syntax to have access to syllable
structure. It is especially strange given that Costa assumes a framework
that involves Late Insertion, such that the actual lexical entries are not
inserted into the syntactic positions until after the syntactic component.
One major downside of this volume is what seems to be the complete lack of
any editorial input on the part of the publisher. It is replete with typos,
inconsistencies in both style and content and spelling mistakes. Although
mostly a minor annoyance, these mistakes can cause confusion or frustration
for the reader. It is often unclear, for example, what difference Costa
assumes between TP and IP (traditionally Tense Phrase and Inflectional
Phrase, respectively). Both terms are found within the same discussion
without any clue as to their specific use. This aspect of the book, of
course, reflects poorly not on the author but on the publisher. The lack of
editing here, unfortunately, seems to be an all too common occurrence these
days in academic publishing. A quick perusal of several recent book reviews
will attest to the growing discontent on the part of the general readership.
Despite the problems mentioned above, I would strongly recommend this book
to anyone interested in any of the topics discussed here, in particular
Romance syntax and the syntax-semantics interface. In brief, this book
makes an important contribution empirically in word order, and topic and
focus in EP and theoretically on syntax and its interface with phonology
and semantics, exploring Reinhart's proposal on the syntax-semantics interface.
Aronoff, Mark. 1994. Morphology by Itself: Stems and Inflectional Classes.
Cambridge, MA: MIT Press.
Bobaljik, Jonathan. 1995. Morphosyntax: the syntax of verbal inflection,
Department of Linguistics and Philosophy, MIT: Ph.D. Dissertation.
Chomsky, Noam. 1995. The Minimalist Program. Cambridge, MA: MIT Press.
Chomsky, Noam. 2001. Derivation by Phase. In Ken Hale: A Life in Language,
ed. Michael Kenstowicz, 1-52. Cambridge, MA: MIT Press.
Di Sciullo, Anna Maria. 2005. Asymmetry in Morphology. Cambridge, MA: MIT
Embick, David, and Noyer, Rolf 2001. Movement Operations after Syntax.
Linguistic Inquiry 32 (4):555-595.
Halle, Morris, and Marantz, Alec 1993. Distributed Morphology and the
Pieces of Inflection. In The View from Building 20: Essays in Linguistic in
Honor of Sylvain Bromberger, eds. Kenneth Hale and Samuel J. Keyser,
111-176. Cambridge, MA: MIT Press.
Julien, Marit. 2002. Syntactic Heads and Word Formation. Oxford: Oxford
Reinhart, Tanya. 1995. Interface Strategies: OTS/Utrecht University.
Reinhart, Tanya. 2006. Interface Strategies: Optimal and Costly
Computations. Cambridge, MA: MIT Press.
Starke, Michal. 2003. Spellout as a trigger for (some) movements, Paper
presented at Workshop on Antisymmetry and Remnant Movement, New York