LINGUIST List 15.1244

Sun Apr 18 2004

Review: Semantics: Landman (2004)

Editor for this issue: Naomi Ogasawara <naomilinguistlist.org>


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 Sheila Dooley Collberg at collberglinguistlist.org.

Directory

  1. Asya Pereltsvaig, Indefinites and the Type of Sets

Message 1: Indefinites and the Type of Sets

Date: Sat, 17 Apr 2004 17:23:28 -0400 (EDT)
From: Asya Pereltsvaig <asya_pereltsvaigyahoo.com>
Subject: Indefinites and the Type of Sets

AUTHOR: Landman, Fred
TITLE: Indefinites and the Type of Sets
SERIES: Explorations in Semantics
PUBLISHER: Blackwell Publishing
YEAR: 2004
Announced at http://linguistlist.org/issues/15/15-567.html


Asya Pereltsvaig, unaffiliated scholar

INTRODUCTION

This book is concerned with the semantics (and to a lesser extent the
syntactic structure) of indefinite nominals, especially (but not
exclusively) those containing numerical expressions, such as 'at least
three girls' or 'the at most five boys'. Although primarily a research
monograph, it can also be used as a text for an advanced/graduate
semantics course. The book presupposes a fairly good knowledge of
formal semantic notation, but the author helpfully identifies which
sections can be read even by people with ''an elementary fluency in
reading semantic types and expressions with lambda-operators'' and
which ones are '''icing' meant for the specialists'' (p. xiii-xv). It
must be also noted that the range of problems considered in this book
makes it interesting and the author's clear and witty style makes it
accessible even for a non-specialist audience, especially for
syntacticians interested in the structure and interpretation of
nominals. The following comments are intended mostly for such a wider
non-specialist audience.

SYNOPSIS

The starting point of this book is the Adjectival Theory of Numerical
Noun Phrases, which the author adopts and further develops into a more
general Adjectival Theory of Indefinite Determiner Phrases
(DPs). Thus, it is argued (Chapters 1 and 2) that indefinite DPs are
generated with the interpretation as the type of sets. Following
Partee (1987), three semantic types are assumed for different types of
nominals: definite DPs (e.g., 'the three girls') are generated with an
interpretation at the type of individuals, <d>; quantificational DPs
(e.g., 'every girl') are generated with an interpretation at the type
of generalized quantifier over individuals, <<d,t>,t>; and indefinite
DPs (e.g., 'three girls', assumed to have a null determiner) are
generated with an interpretation at the type of sets of individuals,
<d,t>. The novelty lies in analyzing indefinites as having the
interpretation which Partee assigns to predicate expressions.

The analysis developed in this book also differs from Partee's
analysis in the kinds of type-shifting operations it allows. Unlike
Partee-style analysis, which allows both lifting and lowering
operations, the Adjectival Theory of Indefinites, in the version
developed in this book, has only lifting operations. Specifically,
Partee derives predicate interpretations of noun phrases from argument
interpretations with type lowering operation BE, which takes a
generalized quantifier <<d,t>,t> and ''maps it onto the set of
individuals for which the property of being that individual is in that
generalized quantifier'' (p. 21). In contrast, the Adjectival Theory
derives argument interpretations of indefinite noun phrases from
predicative interpretations through type lifting with Existential
Closure (EC), an operation which takes ''a set of individuals alpha
and maps it onto a generalized quantifier: the set of all sets that
have a non-empty intersection with alpha'' (p. 21).

The two alternative approaches are compared in Chapter 2, where it is
shown that although the Partee-style approach may have an initial
advantage, both approaches ''must be complex and non-uniform in
analogous ways'' (p. xx). Thus, it is argued that the Classical
(Partee-style) analysis does not win on elegance points. For example,
as pointed out in Chapter 3, both the Classical analysis and the
Adjectival theory need a constraint formulated here as the Variable
Constraint, which says that ''variables cannot be type shifted from
argument types to corresponding predicate types (i.e., from a to
<a,t>)'' (p.49). This constraint in combination with the Adjectival
Theory developed in this book correctly predicts that: (i)
quantificational DPs are infelicitous in predicate position, (ii) DPs
filling predicate position cannot be given wide scope,
(iii)relativization and wh-questioning with the gap in predicate
position are disallowed.

The rest of the book explores the advantages of the Adjectival Theory
in accounting for various empirical phenomena. The first such
phenomenon is that of 'there'-insertion constructions, which is the
topic of Chapters 3 through 8.

The starting point of Chapter 3 is the observation that in addition to
the three restrictions mentioned above for the predicate position,
delayed subjects (also known in syntactic theory as ''thematic
subjects'' or the ''associate'' of the expletive 'there'; e.g., 'a
man' in 'There was a man in the room') exhibit a definiteness effect:
definite DPs are excluded from the delayed subject position. In order
to account for this, Landman analyzes delayed subjects as set-denoting
(on a par with predicates), but crucially as neither predicates, nor
arguments. Instead, delayed subjects are taken (in chapter 4) to be
intersective adjuncts (like intersective adjectives derived from
predicative adjectives generated at the type of sets). Furthermore, it
is proposed that ''a shift into the adjunct domain is only possible
for noun phrases that are generated at the set type, not for noun
phrases that are shifted into the set type'' (p.75). This means that
even though definite DPs can receive a set interpretation in predicate
position (via IDENT operation shifting the interpretation from a to
<a,t>), they cannot receive an interpretation as an intersective
adjunct in delayed subject position. Quantificational DPs are of
course excluded from delayed subject position because they do not have
the right interpretation (i.e., type of sets) to start with.

The rest of Chapter 4 is dedicated to a comparison of this proposal
with alternatives based on a distinction between strong and weak DPs.
Two kinds of theories are considered and shown to be inadequate: those
that identify strength with presuppositionality, and those that
identify weakness with symmetry. The former theories are shown to be
inadequate inasmuch as the correlation between presuppositionality and
infelicity as delayed subject is not that robust; there are
quantificational and definite DPs which are argued not to be
presuppositional but are nonetheless infelicitous as delayed subjects,
and vice versa, there are indefinites which are arguably
presuppositional but are nevertheless felicitous as delayed subjects.
Dutch data concerning *sommige* ('some') and a comparison between
'most boys' and 'more than half of the boys' are used also to refute
the approaches that equate weakness with symmetry.

Chapter 5 is at the core of the book, as it addresses the question of
how DPs in the delayed subject position can be adjuncts if DPs are not
normally licensed as adjuncts. The solution proposed for this problem
consists of two steps. The first step is a modification of the Theta
Theory: the Value Restriction Principle is proposed allowing a
thematic role to be assigned to a constituent not necessarily in an
argument position,but such that ''in the semantic interpretation of
[the minimal syntactic tree] T the interpretation of [that
constituent] constrains the value of role R in the appropriate way''
(p. 104). Hence, the delayed subject is analyzed as an adjunct on the
verbal predicate. The mismatch between the semantic types of the
delayed subject and the verbal predicate is resolved through the
flip-flop shifting mechanism (shifting the interpretation of the
verbal predicate from type <d,<e,t>> to type <e,<d,t>> and vice
versa). This flip-flop mechanism is the technical heart of the book;
it allows a delayed subject, which is syntactically and semantically
an adjunct to receive a thematic role under the Value Restriction
Principle. This flip-flop mechanism also explains why only delayed
subjects are possible, not delayed direct or indirect objects.
Finally, cross-linguistic variation in the realization of non-thematic
subjects in 'there'-insertion constructions in English, German, Dutch
and French is discussed and two parameters are proposed to account for
the four languages: one parameter involves licensing of the null
non-thematic subject and the other involves realization of
non-thematic adverbial (e.g., English 'there', Dutch 'er').

Chapter 6 picks up the question of cross-linguistic variation with
respect to the nature of the predicates that allow adjoined (delayed)
subjects. The flip-flop analysis developed in the previous chapter is
used to explain why adjoined subjects are possible only with
intransitive predicates. Furthermore, it is observed that in English
and French adjoined subjects are possible with unaccusative verbs and
passives and with episodic (i.e., stage-level) predicates, but not
with unergative or transitive verbs or with non-episodic (i.e.,
individual-level) predicates, whereas there is no such restriction in
Dutch and German. The explanation for this contrast is based on a
distinction between saturated and unsaturated predicates. A semantic
parameter is proposed: Dutch and German allow adjunction to all
one-place predicates, whether saturated or unsaturated, whereas
English and French allow adjoined subjects only with saturated
predicates. Bowers' (1993) Pred head is used in order to account for
the position of the delayed subject with respect to the verb in
English and French. A distinction between thematic subject and
reduction subject is used to account for the subject-verb agreement
contrasts between English, German and Dutch, on the one hand, and
French, on the other hand.

Chapter 7 is concerned with 'there'-insertion constructions in Dutch
and argues against semantic partitioning. Chapter 8 is concerned with
negative noun phrases (e.g., 'no girls'), whose appearance as delayed
subjects in 'there'-insertion construction is problematic for many
semantic theories. Here, the semantic break-up approach is pursued,
whereby the negative noun phrase is separated into a negation and an
indefinite noun phrase (i.e., 'not' + 'girls'), and negation takes
scope independently of the rest of the noun phrase. The remaining
three chapters deal with the definiteness effect in other
constructions.

Chapter 9 deals with the definiteness effect with relational DPs
(e.g., 'I have a brother', but not '*I have the brother'). This
definiteness effect is related to the de-thematization of the verb
'have': it loses its possessive meaning and the whole VP 'have a
brother' inherits a relational meaning from the DP. The analysis
proposed for this construction is likewise based on the Adjectival
Theory of Indefinites, and is further extended to constructions with
change of possession verbs, such as 'buy' and 'sell'.

Chapters 10 and 11 deal with the DPs with 'time' occurring in
adverbial position: chapter 10 is concerned with definite 'time'-DPs
(e.g., 'every time the postman came'), while chapter 11 is dedicated
to indefinite 'time'-DPs (e.g., 'three times'). It is argued that
'time' is not a real noun, but a classifier (p. 229). The proposed
analysis takes what looks like a DP in adverbial position to be
(semantically) not a DP, but a perfectly legitimate adverbial
expression.

CRITICAL EVALUATION

On the whole, this book makes an outstanding contribution to the study
of noun phrase meaning and especially definiteness effects. As much as
this book is a comprehensive study on the topic, it is also an
invitation for further research in this area. Many a section of the
book (as the author himself admits) ''requires a syntactic analysis,
rather than presents one'' (p. 118). Here, I will mention just two
such areas of discussion that relate to syntactic controversies.

One involves the relation between predicative and attributive
adjectives. Somewhat confusingly, it is stated on p. 74 that
''predicative adjectives ... are generated at the type of sets... As
adjuncts, they shift to a modifier interpretation''; however, on p. 99
it is stated that ''adjectives... are restricted to adjunct
positions'', creating an impression that predicative adjectives are
derived from attributive ones, rather than vice versa (as in the
previous quote). This relates to the debate in syntactic literature
on whether attributive adjectives are derived from predicative ones or
vice versa. The first approach, deriving attributive adjectives from
predicative ones (typically via a reduced relative clause, e.g., 'a
blond girl' derived from 'a girl who is blond') has been adopted among
others by Cinque (2003) with respect to certain adjectives in
Italian. The second approach, that certain adjectives in predicative
position derive syntactically from attributive adjectives, has been
explored among others by Bailyn (1994), who argues that (long form)
predicative adjectives in Russian are in fact attributive modifiers of
a null Noun.

Another issue that relates to a syntactic controversy is the question
of derivation of OV (i.e., head-final) languages like Dutch and
German. On p. 144, Landman proposes the Adjunction Parameter,
according to which ''DPs with an interpretation born at type <d,t> can
adjoin to one-place predicates'' in Dutch and German, but to
''saturated one-place predicates'' in English and French. On the next
page, this parameter is related to the OV nature of Dutch and German
vs. VO nature of English and French. Furthermore, it is crucially
assumed that ''in Dutch and German the heads in the verbal domain
(like V and I) are sitting on the right side'', that is, that these
languages are underlyingly OV rather than deriving OV order via
movement of the object to the left of the verb. However, this
assumption is not shared by all syntacticians working on this issue
(see papers in Svenonius 2000 for different approaches). It would be
nice to see if the analysis proposed in this book works under a
different set of assumptions, namely that Dutch and German are
underlying VO.

Overall, this book presents an insightful and thorough investigation
of a wide range of natural language phenomena in several languages,
presented in a comprehensible and engaging way. And I cannot but agree
with Veneeta Dayal's comment on the back cover: ''He makes hardcore
linguistics fun to read!''

REFERENCES

Bailyn, John F. (1994) The Syntax and Semantics of Russian Long and
Short Adjectives: An X'-Theoretic Account. In Jindrich Toman (ed.)
Annual Workshop on Formal Approaches to Slavic Linguistics. Ann Arbor,
MI: Michigan Slavic Publications. Pp. 1-30.

Bowers, John (1993) The syntax of predication. Linguistic Inquiry 24:
591-656.

Cinque, Guglielmo (2003) The dual source of adjectives and XP- vs.
N-Raising in the Romance DP. Paper presented at CASTL Kick-Off
conference. University of Tromso.

Partee, Barbara (1987) Noun phrase interpretation and type shifting
principles. In Jeroen Groenendijk, Dick de Jongh and Martin Stokhof
(eds.) Studies in Discourse Representation Theory and the Theory of
Generalized Quantifiers. Dordrecht: Foris.

Svenonius, Peter (ed.) (2000) The Derivation of VO and OV.
Philadelphia: John Benjamins.

ABOUT THE REVIEWER

Asya Pereltsvaig has been investigating issues in the syntax and
interpretation of noun phrases, both in predicative and argument
positions. She is particularly interested in the question of whether
languages without overt articles (such as Slavic languages) employ the
same set of syntactic structures as languages with overt articles. Her
other interests include the syntax of numerical expressions,
interaction of noun phrase syntax with verbal aspect and case marking.
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue