Arguments and adjuncts at the syntax-semantics interface [Elektronische Ressource] / vorgelegt von Roland Schäfer

Informations

• Publié par: georg-august-universitat_gottingen
• Publié le: 01 janvier 2010
• Nombre de lectures: 19

Extrait

Arguments and Adjuncts at the
Syntax-Semantics-Interface
Roland Schäfer M.A.Dissertation
zur Erlangung des Doktorgrades (Dr. phil.) im Fach
Englische Philologie (Neuere englische Sprache),
der
Philosophischen Fakulatät
der Georg-August Universität Göttingen
vorgelegt von
Roland Schäfer M.A.,
geboren am 06. Januar 1974 in Düsseldorf.
Anschrift
Roland Schäfer M.A.
Bundesallee 28
10717 Berlin
Erstgutachter
Prof. Dr. Gert Webelhuth
Seminar für Englische Philologie (Göttingen)
Zweitgutachterin
Prof. Dr. Regine Eckardt
Seminar für Englische Philologie (Göttingen)
Drittgutachter
Priv.-Doz. Dr. Götz Keydana
Sprachwissenschaftliches Seminar (Göttingen)Contents
1 Simpler Semantics 1
1.1 Representational Semantics with Few Types . . . . . . . . . . . . . . . 1
1.1.1 One Basic Type . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1.2 One Type for Arguments and Adjuncts . . . . . . . . . . . . . 3
1.2 Discourse-Level Interpretation . . . . . . . . . . . . . . . . . . . . . . 4
1.3 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2 Event Semantics 7
2.1 Foundations of Event Semantics . . . . . . . . . . . . . . . . . . . . . 7
2.1.1 Arguments for Event-Based Theories . . . . . . . . . . . . . . 8
2.1.2 Roles and Event Individuation . . . . . . . . . . . . . . . . . . 12
2.2 Generalized Operator Approach . . . . . . . . . . . . . . . . . . . . . 16
2.2.1 Operator-Based Approaches . . . . . . . . . . . . . . . . . . . 16
2.2.2 Generalizing the Operator Approach . . . . . . . . . . . . . . . 20
2.2.3 Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.2.4 The Update Procedure . . . . . . . . . . . . . . . . . . . . . . 29
2.2.5 Appendix: Permutation and Drop . . . . . . . . . . . . . . . . 31
3 Quantification 33
3.1 with Quantification . . . . . . . . . . . . . . . . . . . . . . . . . 34
3.1.1 Against Multiple Roles . . . . . . . . . . . . . . . . . . . . . . 34
3.1.2 Simple Quantification . . . . . . . . . . . . . . . . . . . . . . 36
4 Negation, Alternatives, and High Scope 45
4.1 Negation and Event Polarity . . . . . . . . . . . . . . . . . . . . . . . 46
4.1.1 Event Polarity . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
4.1.2 Truth, Falsity, and Updates . . . . . . . . . . . . . . . . . . . . 50
4.2 Types of Negation and Focus . . . . . . . . . . . . . . . . . . . . . . . 55
4.2.1 Basic Distinctions and Phenomena . . . . . . . . . . . . . . . . 55
i
GO
ACONTENTS ii
4.2.2 Focus in . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
4.2.3 Focus and Negation . . . . . . . . . . . . . . . . . . . . . . . . 61
4.2.4 Sentential Negation as Negation and Focus . . . . . . . . . . . 66
4.3 High Scope and Frame Events . . . . . . . . . . . . . . . . . . . . . . 69
4.3.1 Scopal Negation and Some Adverbial Operators . . . . . . . . 70
4.3.2 Where are Frame Events Needed? . . . . . . . . . . . . . . . . 80
4.3.3 Frame Events and Alternatives . . . . . . . . . . . . . . . . . . 86
5 Formalization 88
5.1 Preliminaries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
5.2 Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
5.3 Expressions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
5.3.1 Simple Expressions . . . . . . . . . . . . . . . . . . . . . . . . 90
5.4 The Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
5.4.1 Abstract Models . . . . . . . . . . . . . . . . . . . . . . . . . 93
5.4.2 Secondary Models and Discourse Construction . . . . . . . . . 103
5.5 Inference and Coherence . . . . . . . . . . . . . . . . . . . . . . . . . 105
5.5.1 Consequence . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
5.5.2 Partial and Full Contradiction . . . . . . . . . . . . . . . . . . 113
5.5.3 Undefined Subjects . . . . . . . . . . . . . . . . . . . . . . . . 118
5.6 Sample Derivations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
5.6.1 Lexicalization . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
5.6.2 Logical Forms . . . . . . . . . . . . . . . . . . . . . . . . . . 121
6 Distributivity and Collectivity 126
6.1 Sums and Groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
6.2 Quantifiers, Collectivity, and Distributivity . . . . . . . . . . . . . . . . 130
6.2.1 All Sorts of Readings . . . . . . . . . . . . . . . . . . . . . . . 130
6.2.2 All, Every, and Each, and The Plurals . . . . . . . . . . . . . . 137
7 Implementation within a Syntactic Framework 142
7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
7.2 The HPSG Framework . . . . . . . . . . . . . . . . . . . . . . . . . . 142
7.3 Applicative Semantics in HPSG . . . . . . . . . . . . . . . . . . . . . 143
7.3.1 Semantic Types . . . . . . . . . . . . . . . . . . . . . . . . . . 143
7.3.2 Semantic Values . . . . . . . . . . . . . . . . . . . . . . . . . 145
7.4 Generalized Operator Semantics in HPSG . . . . . . . . . . . . . . . . 148
7.4.1 Lexical Entries . . . . . . . . . . . . . . . . . . . . . . . . . . 148
A
GOCONTENTS iii
7.4.2 Basic Composition . . . . . . . . . . . . . . . . . . . . . . . . 148
7.4.3 Subcategorized NPs . . . . . . . . . . . . . . . . . . . . . . . 150
7.4.4 Proper Names . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
7.4.5 Prepositions and Adverbs . . . . . . . . . . . . . . . . . . . . 153
7.4.6 Scoping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
8 Last Remarks 162
8.1 Achievements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
8.2 Some Major Desiderata . . . . . . . . . . . . . . . . . . . . . . . . . . 163Chapter 1
Simpler Semantics
It is clear that sentences of any natural language have a great deal more
structure than simply the concatenation of one element with another. Thus,
to establish a complexity scale for string sets and to place natural lan-
guages on this scale may, because of the neglect of other important struc-
tural properties, be to classify natural language along an ultimately irrel-
evant dimension. (Partee, ter Meulen and Wall, 1990:436-7)
In this thesis, I define a simplified semantic compositional mechanism based entirely
on Event Semantics, and I provide the skeleton of a flexible syntax-semantics interface
which is formally specified and at the same time open for functional explanations of
grammatical phenomena and pragmatically enriched interpretations. This involves two
major projects: First, I introduce a representational semantic mechanism which does
with a minimal number of semantic types and only one basic type. Secondly, I define
how the semantic representations are interpreted at discourse level to encode and ex-
change information. The project is completed by a proof-of-concept implementation in
a syntactic framework (Head-Driven Phrase Structure Grammar).
1.1 Representational Semantics with Few Types
1.1.1 One Basic Type
Linguists working in formal semantics (a tradition which arguably dates back primarily
to Tarski 1957, but without doubt brought successfully into the linguistic mainstream
by Montague 1973a) and in proof-theoretical frameworks of the syntax-semantics in-
terface like Categorial Grammar (dating back to Ajdukiewicz 1935) usually employ a
1CHAPTER 1. SIMPLER SEMANTICS 2
specific kind of logic (mostly intensional higher-order calculi) and standard model
theory in the semantic analysis of natural language. By making this choice, they set
themselves apart from cognitively oriented linguistic theories and explore the rela-
tions between linguistic expressions and the objects or states of affairs in the material
world rather than mental representations thereof. Every natural language expression is,
from the straightforward viewpoint of formal semantics, a logical formula (in disguise)
which directly receives a disambiguated interpretation in some model.
This highly successful mathematically founded approach has, especially since the 1980s,
been complemented by theories like Discourse Representation Theory (DRT, Kamp
1984, Kamp and Reyle 1993) which provide indirect interpretations for natural lan-
guage expressions by first translating expressions into representations (which could be
but need not necessarily to be similar to mental representations) which then receive an
interpretation at discourse level where concrete models are formed. Especially the loss
of direct interpretation in such frameworks has been criticized by strict model theorists,
a critique which has spawned alternative approaches to the problems solved within
DRT, e.g. in the form of non-representational variants of dynamic logic (Groenendijk
and Stokhof 1990, Groenendijk and Stokhof 1991, Stokhof 2006, etc.).
This thesis follows a representational approach while still keeping up a model-theoretic
primary interpretation of linguistic expressions. How so? Normal model-theoretic se-
mantics is truth-functional. That is, sentences extensionally denote truth-values and are
of the corresponding type t (or Bool or 2). Their truth can be checked in a given model.
The types of the expressions from which the sentence is constructed have to be forged
in a manner that their combination (usually function application) res