LINGUIST List 14.1889

Tue Jul 8 2003

Review: Schwabe and Winkler, ed. (2003)

Editor for this issue: Madhavi Jammalamadaka <>

What follows is a review or discussion note contributed to our Book Discussion Forum. We expect 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 review." Then contact Simin Karimi at


  1. Dimitris Ntelitheos, The Interfaces: Deriving and Interpreting Omitted Structures

Message 1: The Interfaces: Deriving and Interpreting Omitted Structures

Date: Mon, 07 Jul 2003 14:09:32 +0000
From: Dimitris Ntelitheos <>
Subject: The Interfaces: Deriving and Interpreting Omitted Structures

Schwabe, Kerstin and Susanne Winkler, ed. (2003) The Interfaces:
Deriving and Interpreting Omitted Structures, John Benjamins Publishing
Company, Linguistik Aktuell/Linguistics Today 61.

Announced at

Dimitrios Ntelitheos, University of California, Los Angeles

Recent research in elliptical phenomena has shown clearly
that the questions that concern the architectural design of grammar
cannot be answered in strict syntactic, semantic, phonological, or
discourse terms. Thus, the scope of the research has shifted towards
the interfaces between the different grammatical components. Work in
the traditionally investigated interfaces between syntax and semantics
or syntax and phonology has been enriched with additional work on the
interaction between these interfaces and the discourse and information
structural component.

The Interfaces: Deriving and Interpreting Omitted Structures is a
collection of 13 papers that explore the various ways in which
elliptical structures are related to the interfaces of syntax with
semantics, phonology, and discourse.

'Exploring the interfaces from the perspective of omitted structures',
Kerstin Schwabe and Susanne Winkler, 1-26
 In the book's introductory chapter, the editors address general issues
related to current work on ellipsis and the syntax/semantics,
syntax/phonology and syntax/discourse interfaces. They offer a brief
historical background on research in ellipsis and the basic questions that
have been formed in the different research programs throughout the years,
provide some evidence drawn from elliptical phenomena, for the architecture
of the interfaces, and finally present a summary of the different sections
and papers that appear in the collection. The rest of the book is organized
into three parts. The first part is entitled ''Towards the exploration of
PF-deletion Accounts'' and contains four papers that examine phonetic form
(PF) deletion accounts of elliptical structures, a line of thought that was
first presented as the 'phonological reduction hypothesis' in Chomsky and
Lasnik (1993). The second part of the book contains four papers that
investigate elliptic structures from the perspective of the syntax/semantic
interface. The third part contains five papers that explore ellipsis
phenomena from a perspective that concentrates on the relation between
semantics, focus, and discourse structure. The book ends with an extended
section with References (367-387), a Name index (389-393), and a Subject
index (395-399).

The papers are:
 'Ellipsis and syntactic representation', Christopher
Kennedy, 29-53
 Kennedy tries to shed light on the long standing question of whether
constituents targeted by various types of ellipsis-operations have syntactic
structure at some level of representation. It has been known, (see for
example Rooth 1992) that semantic and discourse factors play an important
role in licensing ellipsis. Kennedy shows that a purely syntactic approach
faces certain problems, a fact that has led to the implementation of either
purely semantic approaches or mixed analyses of ellipsis. However, both
semantic and mixed analyses face similar empirical and theoretical problems.
Exploring the interaction of verb phrase (VP)-deletion and parasitic gaps
Kennedy shows that the elided constituent is sensitive to Condition B
effects, strong crossover, and various island constraints (including
wh-islands, Complex noun phrase (NP) Islands, Adjunct Islands and the
Coordinate Structures Constraint). If these are constraints on syntactic
representations, it must then be the case that elided VPs have syntactic
structure. This conclusion is further strengthened by the fact that an
elided VP in these contexts can be shown to require a ''non-parasitic'' gap
analysis in some contexts and a ''missing'' parasitic gap analysis in others.
This is a purely syntactic restriction and thus further supports the fact
that ellipsis constructions are sensitive to configurational constraints on
syntactic representations. Kennedy concludes that a solution lies in
maintaining a purely syntactic approach and explains the problematic cases
by assuming that ''constraints that make reference to the interface between
the syntax and the phonological component should be vacuously satisfied, and
therefore effectively 'turned off''. The idea is that if ellipsis involves
deletion of syntactic structure, then elided constituents should be
sensitive to syntactic constraints in general. However, ellipsis does not
require pronunciation of the omitted structure and so elided constituents
should be insensitive to syntactic constraints that derive from
morpho-phonological properties of lexical items.

'Subject-auxiliary inversion in comparatives and PF output
constraints', Jason Merchant, 55-77
 Merchant's paper is an attempt to explain the fact that I-to-C
movement in comparative clauses can occur only if VP-ellipsis has
deleted the VP complement of the inflection head as in the following
example: 1. a. Abby knows more languages than does her father.
 b. *Abby knows more languages than does her father know. Merchant
proposes that VP-ellipsis in this case is implemented as a repair
mechanism that saves an otherwise illicit structure brought about by
Subject Auxiliary Inversion (SAI); this defect is the ill-formedness
of the intermediate trace of wh-movement that occurs in the
comparative clause. This ill-formedness comes about because all
A'-traces, including the intermediate trace adjoined to VP, are
subject to the Empty Category Principle (ECP) which Merchant
reformulates as a condition that operates at PF. Assuming that the
comparative operator in example 1.b. above, has been extracted from
the object position of 'know', and has moved to specCP by first
adjoining to the VP, I-to-C movement in the comparative clause moves
the auxiliary 'does' out of the IP, forming the chain < does, t
>. While the lower trace satisfies the ECP, I-to-C movement changes
the licensing status of the intermediate trace: the lower copy of
'does', does not PF-head-govern the trace adjoined to VP because it is
not PF-active (i.e. it is not the link at which lexical insertion
occurs; the higher copy in C is). This means that the ECP as
formulated by Merchant is not satisfied. Assuming that VP-ellipsis is
deletion of VP at PF, and because this deletion includes the
offending, intermediate trace of the comparative operator, the ECP is
not violated. Merchant further explores his analysis in three
different contexts: comparatives with overt operators, pseudogapping,
and V-to-I movement, and shows that the empirical facts comply with
his proposal.

'Antecedent-containment and ellipsis', Chris Wilder, 79-119
 Chris Wilder's paper examines antecedent-contained-deletion (ACD)
structures that involve 'wide scope' VP-ellipsis as in the following
example: 2. John said that more trees had died than Mary did In 2, the
comparative clause containing the elided VP is itself contained in a
CP inside the VP that antecedes the ellipsis. These structures seem to
be possible only when the comparative clause is extraposed within this
complement CP in surface order. Wilder shows that previous accounts
(i.e. quantifier raising, extraposition, and A-movement) face
empirical problems. He shows that the traditional constraint that
prohibits antecedent- containment at LF is too liberal to account for
the distribution of well-formed and ill-formed ACDs and that a further
constraint is needed, which prohibits 'PF-containment' of the ellipsis
site by its antecedent. Apparent counterexamples to this proposal are
analyzed as resulting from the interaction of two independent ellipsis
rules; pseudogapping and backward deletion and thus the elliptical
sites are claimed not to be antecedent-contained.

'Background matching in right node raising constructions', Katharina
Hartmann, 121-151
 Hartmann's paper departs from the above three approaches in that it
formulates a PF-deletion account of ellipsis that imposes a pragmatic
condition that becomes the licensing condition of ellipsis at PF. The
main focus of the paper is right-node raising (RNR) constructions in
German and in particular the investigation of the distribution of
accents and its consequences on the focus structure of RNR
structures. Evidence shows that RNR does not have to be a syntactic
constituent as it can strand prepositions and violates
islands. Hartmann shows that the elements that immediately precede the
targets of RNR in both conjuncts must contrast. Using Schwarzschild's
(1999) theory of GIVENess and an extension of the notion of discourse
antecedents the paper shows that the targets of RNR serve as discourse
antecedents of the other conjunct resulting in deaccenting. The
distribution of accent and the phonetic identity of the targets are
shown to be obligatory conditions for the 'Principle of Pragmatic
Licensing' which states that an utterance is pragmatically licensed if
it has a background match, the latter being defined as GIVENess.
Thus, PF deletion is the result of the interaction between information
structure, a parallel syntactic configuration and a specific
intonational pattern in RNR.

'Merge copy', Caterina Donati, 155-175
 Donati argues that both a PF and an LF process are needed to
account for ellipsis but the basic mechanism underlying ellipsis is
neither phonological nor semantic, but purely syntactic. As all
syntactic phenomena ellipsis gets interpreted at both interfaces but
it is not in itself an interface process. She proposes a new
mechanism termed ''merge copy'' which forms part of the definition of
Move. Given a certain numeration and having constructed K by merging a
and b, it is possible to merge K with a copy of b. Under this
perspective, the deletion effect of ellipsis in PF becomes clear: it
is simply an instance of the more general mechanism of ''delete
copy''. Thus ellipsis in general is reduced to movement. The proposed
process has two distinct instantiations one in standard movement
operations and one in reduplication (i.e. ellipsis phenomena). The
main difference is that in standard movement the copies are links of
the same chain while in reduplication the copies are members of
different chains with the result that no agreement relation holds
between them. Donati also discusses the problem that the optionality
of ellipsis poses for a movement analysis and proposes that this
optionality can be accounted for by assuming different enumerations
for the two choices.

'Phrase structure paradoxes, movement and ellipsis', Winfried Lechner,
 Lechner argues that the systematic differences between traces and ellipsis
copies (for example their varying ability to host reconstruction sites for
movement) do not reveal intrinsic properties of the two different exponents
of copies but can be derived from general principles of economy. Evidence
drawn from ellipsis and movement phenomena shows that there are derivations
in which a category can be potentially merged into two distinct locations
and that the choice between the competing candidates is determined by
economy. The empirical facts in the paper come from two types of
constructions in which a phonetically silent VP is followed by an overt
remnant: so-called Phrase Structure Paradoxa involving VP-fronting and
instances of VP-Ellipsis or pseudogapping. Lechner proposes a movement
analysis of PF-paradoxes assuming extraction of the remnant PP prior to
topicalization. Thus, VP-fronting receives the same analysis as
pseudogapping (Johnson, 1996). The interpretive differences between the two
phenomena are explained via the assumption that VP adjuncts may be merged in
any position in which they are interpretable and that these positions are
determined by economy conditions. The consequence of this is that economy is
a factor that restricts both structure-building operations of movement and

Unpronounced heads in relative clauses Uli Sauerland 205-226
 Sauerland investigates English relative clauses. The basic
assumption is that a satisfactory analysis of relative clauses should
posit two different sources for the head: a clause-internal source as
in Kayne (1994) and a base-generated source. Sauerlnd argues further
that the structure of the base-generated relative clause further
involves a silent copy of the head in the corresponding
clause-internal position (i.e. a lower copy in a PF chain) that is
deleted. Using diagnostics for wh-movement and especially
reconstruction effects of movement Sauerland shows that it is indeed
the case that both types of relative clauses must exist. Following the
standard assumption that 'vehicle change' is possible in ellipsis but
not in movement chains, he shows that the silent copy of the head is
related to the overt copy not by movement but by ellipsis and
specifically a process that he calls 'relative deletion', an operation
similar to comparative deletion

'Variation at the syntax-semantics interface: Evidence from gapping',
Luis López and Susanne Winkler, 227-248
 As the title indicates, the main purpose of López and Winkler's
paper is to account for cross-linguistic variation in wh and focus
movement structures within the Minimalist Program. Some languages
(Bade, Aghem, Hungarian) move wh/focus-phrases to a clause internal
position, which can be identified as Spec v. This may be the case for
English as well. Johnson (1996) has argued that the second conjunct of
a gapping construction is a vP. Following his proposals López and
Winkler, analyze gapping as vP coordination plus across-the-board
movement. The investigation of the properties of the movement process
provides support to the claim that the focused remnants of gapping
must occur in spec-vP in English. If this argument is on the right
track, then it is not surprising that wh/focus-phrases exist in
English gapping or topicalized constructions. López and Winkler
propose that cross-linguistic variation is the result of different
cross-linguistic interpretive rules and more crucially that these
interpretive rules are ranked. Variation is the result of the
alternative rankings of those rules.

'Ellipsis and the structure of discourse', Daniel Hardt, 251-262
 Hardt argues that the interpretation of ellipsis is subject to
constraints based on the structure of discourse. He makes two basic
claims: ellipsis resolution requires that a matching relation holds
between a containing clause and some antecedent clause (Rooth, 1992),
and that clauses in discourse are structured according to discourse
relations while ellipsis resolution occurs as a side effect of
establishing these discourse relations. He considers cases problematic
for purely semantic analyses of ellipsis such as the existence of
multiple potential antecedents for VP-ellipsis, the ''many clause
puzzle'' in which two ellipsis occurrences are preceded by a single
antecedent clause, as well as other multiple-ellipsis data, and shows
that semantic matching must be applied according to discourse

'Correlate restriction and definiteness effect in ellipsis', Maribel
Romero, 263-300
 This paper is concerned with two ellipsis constructions: Reduced
Conditionals (in German) and Sluiced Interrogatives Clauses (in
English and many other languages), illustrated in 3.a. and
3.b. respectively: 3. a. Wenn ich wen besuche, dann (immer) den Peter.
 If I somebody visit then always the Peter
 ''If / whenever I visit somebody, then Peter / it's Peter.''
 b. Somebody just left - guess who.

 Romero shows that these structures exhibit two peculiar
characteristics that make them look rather different from other types
of ellipsis: the restriction on possible antecedent phrases for the
remnants of ellipsis, and a definiteness effect that makes
non-definite phrases behave semantically as definites in ellipsis
sites. Romero argues against current approaches in the literature that
posit an idiosyncratic account for each of these types of ellipsis and
shows that these facts follow from the interaction of fairly standard
assumptions about ellipsis. These include the presence of Focus in the
remnant material and the semantics of conditionals and questions.

'F-marking and specificity in sluicing constructions', Kerstin
Schwabe, 301-319
 Schwabe presents a novel analysis of sluicing constructions
investigating two factors that play a crucial role in the licensing of
sluicing: a focus restriction based on Schwarzschild (1999) and a
specificity restriction. She argues that the wh-phrase in the sluicing
sentence and the related phrase in the antecedent clause must be
F-marked. Furthermore, the relatum must be an indefinite that allows
for a specific interpretation, specificity being an anchoring relation
between the discourse referent and a discourse given item. Since
specific indefinite elements present new information (are not GIVEN)
sluicing cannot be licensed in certain contexts including the scopal
domain of definite DPs, the scopal domain of thematic matrix
predicates, and of downward-monotone quantifiers. All these contexts
exhibit non-novel indefinites.

'The semantics of Japanese null pronouns and its cross-linguistic
implications', Satoshi Tomioka, 321-339
 Tomioka explores the semantics of silent pronouns in Japanese. He
says that these elements receive a surprisingly wide variety of
semantic interpretations. The paper presents an analysis of these
pronouns as a phonologically null version of bare NPs that requires a
short number of semantic operations for its interpretation. An
examination of cross-linguistic data shows that Tomioka's proposal
makes correct predictions on the semantic variability of null
arguments in other languages. It is predicted that languages like
Japanese that allow for both bare NP arguments and null pronouns will
allow for the latter to exhibit a semantic variability closely tied to
the variability of NP interpretation in general. This is shown to hold
for Chinese and Korean. On the other hand languages that employ only
one of the introduced semantic tools will allow only a certain type of
nominal expressions to go phonologically null. Greek seems to fit this

'Omission impossible? Topic and focus in Focal Ellipsis', Petra
Gretsch, 341-365
 Gretsch's paper examines the interdependencies of focus-structure
and topic-interpretation by analyzing focal ellipsis (FE). Her main
thesis states that topic-interpretation (in German) is exclusively
dependent upon the syntactic focus structure of a sentence which is
responsible for the information structuring. Neither accent-driven
accounts nor topological-driven accounts (relying on a functional
topic position) are necessary. Moreover, the commonalities plus
differences between internal and external topics fall out for
free. Gretsch differentiates between two cases: presentational
vs. contrastive focus-omission, with the former having one and the
latter two focus domains. The paper further argues that FE are not
hidden wh-questions - contrary to their function - but declarative
structures with a special gap: the focus (exponent) is
missing. Gretsch uses data from German, Chinese, and Korean, to
illustrate these points. Thus, Chinese FE with omission of the
contrastive focus have to exhibit the pragmatically induced
sentence-particle '-ne' whereas presentational FE don' t allow for
that particle. Finally, the paper argues for a syntactic analysis of
German without the functional projections 'TopicPhrase' and
'FocusPhrase'. Instead Gretsch assumes an information-structural topic
interpretation which relies on the syntactic focus-structure as
exclusive topic-indicator.


This volume is a valuable contribution to the study of ellipsis and in
particular the role that the interaction of different grammatical
components play in the licensing of elliptical structures. The
contributions to the volume cover a wide range of theoretical problems
that go beyond the traditionally investigated interfaces between
syntax and semantics or syntax and phonology. Additional work on the
interaction between these interfaces and the discourse and information
structural component has received special focus in line with current
approaches in related research. Three equally important aspects of
ellipsis have been covered in detail. The long-standing problem of
whether ellipsis is a PF deletion process is covered in the first part
of the book (Kennedy, Merchant, Wilder, and Hartmann). The authors
present data from different elliptical phenomena (comparative
deletion, antecedent-contained deletion and right-node raising) and
argue convincingly for a PF deletion account of the ellipsis.

A bolder syntactic approach is implemented in the second part of the
book. The papers here (Donati, Lechner, Sauerland, López & Winkler)
try to reduce ellipsis to syntactic movement. If this is on the right
track it will definitely be a desired consequence, as it would
drastically simplify Universal Grammar. The main argument against such
approaches comes from the fact that ellipsis and movement exhibit
different properties with respect to island constraints. These ideas
have precedents in for example, Johnson (2001) where VP Ellipsis is
shown to be subject to licensing conditions that recall conditions on
traces, ultimately suggesting that we should derive VP Ellipsis by way
of movement. Lechner in the present volume, shows that this is
actually the case, i.e. that VP-fronting and VP-ellipsis are
essentially the same and that any differences in their distribution
can be accounted for if we assume certain principles of economy in the
syntactic derivation. Finally, López & Winkler's paper is also based
on Johnson's work in gapping, assuming an across the board (ATB)
movement of the verb and not deletion. This is taken as a starting
point for a successful investigation of linguistic variation.

The third part of the book is dedicated to an approach towards
ellipsis, that explores semantic and discourse aspects of the
phenomena involved. The approaches in this part follow a line of
research that relies heavily on Rooth (1992) and Schwarzschild, (1999)
and the association of ellipsis to semantic focus restrictions. The
problem with the syntactic approaches is that they seem to be unable
to account for a number of puzzles (i.e. the many-clause puzzle
investigated in Hardts paper, specificity effects in sluicing in
Schwabe, semantic diversity in the interpretation of null pronouns in
Tomioka). Most of the semantic approaches towards ellipsis associate
the licensing mechanism of ellipsis to some sort of focus condition
that licenses the omission of backgrounded material. However, Gretsh's
paper shows that this is not always true. She provides very
interesting data from German that shows that ellipsis of focused
material is also possible and that the semantic approaches need to be
modified in order to capture this type of ellipsis too.

As the editors state in their introduction, the goal of the volume is
to present an overview of the current state of the art in research of
ellipsis and omitted elements. As far as work in the interfaces is
concerned this has been achieved with the collection of papers
included in the volume. However, there are alternative approaches
towards ellipsis that receive no attention at all in any of the three
parts of the volume. In the domain of syntax for example, Lobeck
(1995, and subsequent work) followed by a number of different
researchers has argued equally successfully that ellipsis involves a
null pronominal element (pro) explaining the pronoun-like properties
that VP-ellipsis and N'-Drop exhibit. The discussion would have been
benefited if a related approach were included in this first part of
the book. Obviously the inclusion of every available proposal on
ellipsis cannot be achieved in a collection of this
size. Consequently, the final table of contents is an excellent
representation of the diversion of approaches towards elliptical
phenomena cross-linguistically.


Chomsky, Noam & Lasnik, Howard (1993) The Theory of Principles and
Parameters, in J. Jacobs et al (eds.) Syntax: An International
Handbook of Contemporary Research [Volume 1], Berlin and New York:
Walter de Gruyter, 506-569.

Johnson, Kyle (1996) In Search of the English Middle Field,
unpublished manuscript.

Johnson, Kyle (2001) What VP ellipsis can do, what it can't, but not
why, in M. Baltin and C. Collins (eds.) The handbook of contemporary
syntactic theory, Blackwell Publishers, pp. 439-479.

Kayne, Richard (1994) The Antisymmetry of Syntax. Cambridge, Mass: MIT

Lobeck, Anne (1995) Ellipsis: Functional Heads, Licensing and
Identification. New York, Oxford University Press.

Rooth, Mats (1992) Ellipsis redundancy and reduction redundancy, in S.
Berman and A. Hestvik (eds.) Proceedings of the Stuttgarter Ellipsis
Workshop, Arbeitspapiere des Sonderforschungsbereichs 340, 29.

Schwarzschild, Roger (1999) GIVENness, AVOIDF, and Other Constraints
on the Placement of Accent, Natural Language Semantics 7, 141-177.


Dimitris Ntelitheos is a graduate student at the Department of
Linguistics, University of California, Los Angeles. His academic
interests include nominal ellipsis, discontinuity in the DP, word
order variation, adjectival syntax and other DP-internal syntactic
phenomena in Greek and other languages.
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue