Introduction to Semantic Theory Modelling the context: Pronouns

Similar documents
Proof Theory for Syntacticians

Unit 8 Pronoun References

Focusing bound pronouns

Parsing of part-of-speech tagged Assamese Texts

A Minimalist Approach to Code-Switching. In the field of linguistics, the topic of bilingualism is a broad one. There are many

Introduction to HPSG. Introduction. Historical Overview. The HPSG architecture. Signature. Linguistic Objects. Descriptions.

Compositional Semantics

Constraining X-Bar: Theta Theory

Using dialogue context to improve parsing performance in dialogue systems

Chapter 4: Valence & Agreement CSLI Publications

Dear Teacher: Welcome to Reading Rods! Reading Rods offer many outstanding features! Read on to discover how to put Reading Rods to work today!

Segmented Discourse Representation Theory. Dynamic Semantics with Discourse Structure

Inleiding Taalkunde. Docent: Paola Monachesi. Blok 4, 2001/ Syntax 2. 2 Phrases and constituent structure 2. 3 A minigrammar of Italian 3

Linguistic Inquiry, Volume 35, Number 1, Winter 2004, pp (Article)

Context Free Grammars. Many slides from Michael Collins

ELD CELDT 5 EDGE Level C Curriculum Guide LANGUAGE DEVELOPMENT VOCABULARY COMMON WRITING PROJECT. ToolKit

ENGBG1 ENGBL1 Campus Linguistics. Meeting 2. Chapter 7 (Morphology) and chapter 9 (Syntax) Pia Sundqvist

Types and Lexical Semantics

Universal Grammar 2. Universal Grammar 1. Forms and functions 1. Universal Grammar 3. Conceptual and surface structure of complex clauses

Syntax Parsing 1. Grammars and parsing 2. Top-down and bottom-up parsing 3. Chart parsers 4. Bottom-up chart parsing 5. The Earley Algorithm

Theoretical Syntax Winter Answers to practice problems

CS 598 Natural Language Processing

Dual Content Semantics, privative adjectives, and dynamic compositionality

Some Principles of Automated Natural Language Information Extraction

Approaches to control phenomena handout Obligatory control and morphological case: Icelandic and Basque

AN EXPERIMENTAL APPROACH TO NEW AND OLD INFORMATION IN TURKISH LOCATIVES AND EXISTENTIALS

Case government vs Case agreement: modelling Modern Greek case attraction phenomena in LFG

Intra-talker Variation: Audience Design Factors Affecting Lexical Selections

Basic Parsing with Context-Free Grammars. Some slides adapted from Julia Hirschberg and Dan Jurafsky 1

1/20 idea. We ll spend an extra hour on 1/21. based on assigned readings. so you ll be ready to discuss them in class

Aspectual Classes of Verb Phrases

English for Life. B e g i n n e r. Lessons 1 4 Checklist Getting Started. Student s Book 3 Date. Workbook. MultiROM. Test 1 4

Language acquisition: acquiring some aspects of syntax.

Argument structure and theta roles

Agree or Move? On Partial Control Anna Snarska, Adam Mickiewicz University

Grammars & Parsing, Part 1:

a) analyse sentences, so you know what s going on and how to use that information to help you find the answer.

Writing a composition

Underlying and Surface Grammatical Relations in Greek consider

Engineers and Engineering Brand Monitor 2015

Derivational and Inflectional Morphemes in Pak-Pak Language

Procedia - Social and Behavioral Sciences 141 ( 2014 ) WCLTA Using Corpus Linguistics in the Development of Writing

Control and Boundedness

Construction Grammar. University of Jena.

BULATS A2 WORDLIST 2

Words come in categories

Language Acquisition Fall 2010/Winter Lexical Categories. Afra Alishahi, Heiner Drenhaus

Managerial Decision Making

CSC200: Lecture 4. Allan Borodin

"f TOPIC =T COMP COMP... OBJ

Natural Language Processing. George Konidaris

The Effect of Discourse Markers on the Speaking Production of EFL Students. Iman Moradimanesh

L1 and L2 acquisition. Holger Diessel

Corpus Linguistics (L615)

Written by: YULI AMRIA (RRA1B210085) ABSTRACT. Key words: ability, possessive pronouns, and possessive adjectives INTRODUCTION

Author: Fatima Lemtouni, Wayzata High School, Wayzata, MN

SAMPLE PAPER SYLLABUS

AN ANALYSIS OF GRAMMTICAL ERRORS MADE BY THE SECOND YEAR STUDENTS OF SMAN 5 PADANG IN WRITING PAST EXPERIENCES

cambridge occasional papers in linguistics Volume 8, Article 3: 41 55, 2015 ISSN

Derivational: Inflectional: In a fit of rage the soldiers attacked them both that week, but lost the fight.

Intension, Attitude, and Tense Annotation in a High-Fidelity Semantic Representation

Basic Syntax. Doug Arnold We review some basic grammatical ideas and terminology, and look at some common constructions in English.

The Agile Mindset. Linda Rising.

THE SHORT ANSWER: IMPLICATIONS FOR DIRECT COMPOSITIONALITY (AND VICE VERSA) Pauline Jacobson. Brown University

Prediction of Maximal Projection for Semantic Role Labeling

Phenomena of gender attraction in Polish *

Enhancing Unlexicalized Parsing Performance using a Wide Coverage Lexicon, Fuzzy Tag-set Mapping, and EM-HMM-based Lexical Probabilities

Visual CP Representation of Knowledge

Objectives. Chapter 2: The Representation of Knowledge. Expert Systems: Principles and Programming, Fourth Edition

THE INTERNATIONAL JOURNAL OF HUMANITIES & SOCIAL STUDIES

Discussion Data reported here confirm and extend the findings of Antonucci (2009) which provided preliminary evidence that SFA treatment can result

CORPUS ANALYSIS CORPUS ANALYSIS QUANTITATIVE ANALYSIS

Anaphoric Resolution of Singular and Plural Pronouns: The Reference to Persons Being Introduced by Different Co-ordinating Structures

Morphosyntactic and Referential Cues to the Identification of Generic Statements

Type Theory and Universal Grammar

Probabilistic Latent Semantic Analysis

Developing Grammar in Context

CAAP. Content Analysis Report. Sample College. Institution Code: 9011 Institution Type: 4-Year Subgroup: none Test Date: Spring 2011

(Sub)Gradient Descent

Short Text Understanding Through Lexical-Semantic Analysis

Lesson 2. La Familia. Independent Learner please see your lesson planner for directions found on page 43.

Controlled vocabulary

The Structure of Multiple Complements to V

CAS LX 522 Syntax I. Long-distance wh-movement. Long distance wh-movement. Islands. Islands. Locality. NP Sea. NP Sea

(3) Vocabulary insertion targets subtrees (4) The Superset Principle A vocabulary item A associated with the feature set F can replace a subtree X

THE SOME INDEFINITES

Monitoring Metacognitive abilities in children: A comparison of children between the ages of 5 to 7 years and 8 to 11 years

Switched Control and other 'uncontrolled' cases of obligatory control

The subject of adjectives: Syntactic position and semantic interpretation

The College Board Redesigned SAT Grade 12

Transitive meanings for intransitive verbs

Learning Methods in Multilingual Speech Recognition

DESIGNING NARRATIVE LEARNING MATERIAL AS A GUIDANCE FOR JUNIOR HIGH SCHOOL STUDENTS IN LEARNING NARRATIVE TEXT

MODELING DEPENDENCY GRAMMAR WITH RESTRICTED CONSTRAINTS. Ingo Schröder Wolfgang Menzel Kilian Foth Michael Schulz * Résumé - Abstract

Sample Goals and Benchmarks

I. INTRODUCTION. for conducting the research, the problems in teaching vocabulary, and the suitable

Case study Norway case 1

The Role of the Head in the Interpretation of English Deverbal Compounds

Semantic Inference at the Lexical-Syntactic Level for Textual Entailment Recognition

An Approach to Polarity Sensitivity and Negative Concord by Lexical Underspecification

Transcription:

Introduction to Semantic Theory Modelling the context: Pronouns Class: June 15, 2016

Recap and aim Connecting back to the previous lecture Central result: to account for attributive adjectives, we extended the formal system with the derivation rule (PM) We derived a proposal for the semantic type and the extension of the definite article. We discussed two possible solutions to the empirical problem provided by attributive adjectives. We discussed their conseqences, and chose to introduce a new derivation rule predicate modification (PM). (1) Predicate Modification (PM): For a branching node α with the set of daughters {β, γ}, where β and γ are of type e, t, then α w = λx e.β w (x) & γ w (x)

Recap and aim Aim for today The aim for today: to discuss expressions that heavily depend on context: personal pronouns. Personal pronouns have variable interpretations. Their kind of variability cannot be captured by the inherently static extensions that we used before. Consequence: extension of the system by formal means to capture the context dependence of personal pronouns. We will see: third person and first/second person singular pronouns have different semantic behavior and need to be analyzed differently. Methodological lesson: how to get a better picture of the semantic contribution of an expression α by contrasting and comparing it with (superficially) similar expressions.

The data Individual denoting expressions We have seen two types of linguistic expressions that denote an individual: proper names definite descriptions (the+np) They have in common: they are expressions of type e Their semantic difference lies in how the denotation to a specific individual is fixed.

The data Proper names vs. definite descriptions Given the extensions we determined, what is the semantic difference between the two types of expressions? (2) a. Peter w = Peter b. the boy w = ιx[boy (x)(w)]

The data Proper names vs. definite descriptions Given the extensions we determined, what is the semantic difference between the two types of expressions? (2) a. Peter w = Peter b. the boy w = ιx[boy (x)(w)] For proper names, the individual that is denoted is fixed as part of their lexical entry. For definite descriptions, the individual that is denoted is fixed by the definite article on the basis of the extension of the noun and the context of utterance (this is done by the ι-operator).

The data Another class of expressions denoting individuals: pronouns At first we will restrict ourselves to third person singular pronouns, he, she, it (and their case variants). (3) a. Peter told Paul a story that he liked. b. Peter told Paul a story that Peter liked. c. Peter told Paul a story that Paul liked. (4) Peter told Paul a story that she liked. Describe what the examples in (3) and (4) show regarding the extensions of third person singular pronouns in contrast to proper names!

The data The three types of uses of 3rd sg pronouns As examples (3) and (4) suggested, third person singular pronouns have variable denotation. They can denote varying individuals that fit their gender specification. Three different uses are traditionally distinguished. Referential anaphoric pronouns: (5) John called. He ll drop by later.

The data The three types of uses of 3rd sg pronouns As examples (3) and (4) suggested, third person singular pronouns have variable denotation. They can denote varying individuals that fit their gender specification. Three different uses are traditionally distinguished. Referential anaphoric pronouns: (5) John called. He ll drop by later. Deictic (referential) pronouns: (6) [speaker points at John:] He did it!

The data The three types of uses of 3rd sg pronouns As examples (3) and (4) suggested, third person singular pronouns have variable denotation. They can denote varying individuals that fit their gender specification. Three different uses are traditionally distinguished. Referential anaphoric pronouns: (5) John called. He ll drop by later. Deictic (referential) pronouns: (6) [speaker points at John:] He did it! Bound anaphoric pronouns: (7) No actress resents her cat.

The data How to model 3rd sg pronouns? I Until this point, we have only seen expressions that always contribute (qualitatively!) the same extension. (8) tree w = λx e.tree (x)(w) Even though the trees that are in this set vary depending on what the world w looks like, the noun tree always denotes the set of all trees in w (or the corresponding function). The variable denotation of personal pronouns is conceptually different. The pronoun he does not always denote the same male individual in w not even in a sequence of utterances. (9) Peter called. He ll drop by later to talk to Paul. He had asked for him.

The data How to model 3rd sg pronouns? II Given the discussion above, can we assign an extension to third person singular pronouns in the same manner as we did for noun, verbs, etc?

The data How to model 3rd sg pronouns? II Given the discussion above, can we assign an extension to third person singular pronouns in the same manner as we did for noun, verbs, etc? No, this is not possible. The variability just cannot be captured by the relative invariability of fixed extensions. Goal: We need to augment the model by something that can provide different individuals as denotations of these pronouns.

Enter: assignment functions Individuals provided by the (discourse) context For the deictic and the referential anaphoric uses of third person singular pronouns, we can argue that the individuals are provided either by the utterance context (deictic use) or the previous discourse (anaphoric use). (10) a. [speaker points at John:] He did it! b. John called. He ll drop by later. We need something that makes individuals from the utterance context accessible and that tracks individuals introduced in the previous discourse.

Enter: assignment functions Idea: assignment functions I To make individuals from the utterance context accessible and to track individuals introduced in the previous discourse, we introduce assignment functions into the model. (11) An assignment function g is a function from indices to individuals. This means, it assigns to each index one and only one individual. (12) Indices are modelled by natural numbers (1, 2, 3,...) that are assigned to expressions that denote individuals. In case two expressions share the same index, they denote the same individual. John 1 called. He 1 ll drop by later.

Enter: assignment functions Idea: assignment functions II The example illustrating the referential (anphoric) use suggests that proper names introduce conditions on the assignment function g, and pronouns access g to get their interpretation. To formalize this, we add g to the interpretation function. This is meant to say that any expression that accesses the assignment function, accesses this g. (13) Mary 5 w,g = Mary [Condition on g: g(5) = Mary]

Enter: assignment functions Idea: assignment functions III In the deictic use, the pointing gesture seems to fix an individual as the denotation of a pronoun. To make this case parallel to the referential (anaphoric) case, we will still assume that the gesture places a restriction (like proper names) on the assignment function g for whichever index the pronoun carries. (14) [speaker points at John:] He 2 did it! The pointing introduces the following condition on g: g(2) = John

Interpreting 3rd sg pronouns Interpreting pronouns For the referential anaphoric pronouns and the deictic pronouns, we want that their denotation is the individual that the assignment function assigns to the index that the pronoun carries. Mary 5 went to a party. She 5 had a good time. Condition: g(5) = Mary Assign: g(5) = Mary

Interpreting 3rd sg pronouns Introducing a new rule Since the contribution of pronouns is not fixed for w, but depends on the make-up of g, we cannot formulate the extension of a pronoun in the same manner as we did for all other lexical items before. In fact, the only way to formulate the variable interpretation is via an interpretation rule. (15) Pronouns Rule (PR): If α is a pronoun, then for any assignment g and index i for which g assigns a value: α i w,g = g(i) Which type does the result of the new interpretation rule have? That is, what is the type of pronouns?

Interpreting 3rd sg pronouns Sample derivation Assume: a previous utterance established g(1) = Mary. DP e She 1,e VP t V e,t laughs e,t w,g 2 (NN) = She 1,e VP t laughs e,t (FA) = laughs w,g (she 1 w,g ) (PR) = [λx. laugh (x)(w)](g(1)) = [λx. laugh (x)(w)](mary) λ = 1 iff laugh (Mary)(w) w,g

Interpreting 3rd sg pronouns What about the bound use of pronouns? Before we can say anything substantial about this use, we need a hypothesis for the extension of quantifying expressions ( later classes). (16) Every actress loves her cat. Intuitively, the interpretation of her in (16) seems to covary with the actresses that the quantifying expression varies over: each actress has a cat and she loves that cat.

Interpreting 3rd sg pronouns Interim summary I Like proper names and definite descriptions, third person singular pronouns denote individuals (type e). Third person singular pronouns do not have fixed extensions, though. Their denotations are determined via the Pronouns Rule (PR): (17) If α is a pronoun, then for any assignment g and index i for which g assigns a value: α i w,g = g(i) (PR) utilizes an assignment function g which stores individuals that are contextually given under different indices; it accesses g to pick out the interpretation of a pronoun.

Interpreting 3rd sg pronouns Interim summary II Linking back to the classes on ambiguities: based on how we model personal pronouns, how would we formulate formally why for certain sentences referential ambiguities arise? (18) Peter told Paul that he should call Mary.

Interpreting 3rd sg pronouns Interim summary II Linking back to the classes on ambiguities: based on how we model personal pronouns, how would we formulate formally why for certain sentences referential ambiguities arise? (18) Peter told Paul that he should call Mary. The referential ambiguities arise in case a referentially used (anaphoric) pronoun, e.g., he in (18), can in principle be co-indexed with more than one antecedent above, either Peter or Paul: (19) a. Peter 1 told Paul 2 that he 1 should call Mary. b. Peter 1 told Paul 2 that he 2 should call Mary.

The data First and second person singular pronouns I First person singular (I) and second person singular (you) pronouns are traditionally grouped with third person singular pronouns (he, she, it) and all plural personal pronouns (which we will ignore). Given this grouping, the obvious question is: Do first and second person singular pronouns behave semantically just like third person singular pronouns? And formally, can we account for their behavior with the same mechanisms that we introduced for third person singular pronouns?

The data First and second person singular pronouns II Consider the following examples containing first person singular (I) and second person singular (you) pronouns: (20) A: What did you do last weekend? B: I was at the department and prepared my lecture for today. A: Really? I watched a lot of TV! Which use of third person singular pronouns shares a similar behavior to that of I and you in (20)?

The data Data comparison I Third person singular pronouns in their deictic use behave differently from first and second person singular pronouns: third person pronouns can be used anaphorically to denote the same individual as another pronoun in a previous discourse turn. (21) A: [points at Peter] He was at the movies yesterday. B: Really? I thought he had the flu!

The data Data comparison I Third person singular pronouns in their deictic use behave differently from first and second person singular pronouns: third person pronouns can be used anaphorically to denote the same individual as another pronoun in a previous discourse turn. (21) A: [points at Peter] He was at the movies yesterday. B: Really? I thought he had the flu! (22) A: I was at the movies yesterday. B: #Really? I thought I had the flu! (23) A: Did you go to the movies yesterday? B: #No, you had the flu.

The data Data comparison II First and second person singular pronouns also behave differently from proper names and definite descriptions: they can also denote the same individual as another occurrence of the same expression in another person s utterance. (24) Peter: Mary is nice. A to B: Peter said that Mary was nice. (25) Peter: The girl is nice. A to B: Peter said that the girl was nice.

The data Data comparison II First and second person singular pronouns also behave differently from proper names and definite descriptions: they can also denote the same individual as another occurrence of the same expression in another person s utterance. (24) Peter: Mary is nice. A to B: Peter said that Mary was nice. (25) Peter: The girl is nice. A to B: Peter said that the girl was nice. (26) Peter: I m nice. A to B: #Peter said that I was nice.

The data Differences: how the interpretation is fixed Proper names: stable, fixed denotation independent of the utterance context Definite descriptions: the individual that is denoted is stable in an utterance context Third singular pronoun: the individual that is denoted is stable in an utterance context (as long as the occurrences bear the same index and the assignment function is not changed) First and second singular pronouns: the individual that is denoted shifts with every discourse turn

Interpreting 1st and 2nd sg pronouns First and second person singular and the utterance context The interpretation of first person singular (I) and second person singular (you) is intimately tied to the utterance context since the individuals that are denoted shift at every turn i.e., whenever the speaker changes. In contrast to third person singular pronouns the shift/fixing of the individual is automatic and does not depend on speaker intentions: (27) A: [points at Peter] He likes Mary. A : [points at Peter] #I like Mary.

Interpreting 1st and 2nd sg pronouns Formalizing I and you The formalization that we will adopt was proposed by Kaplan (1977), who introduced a context parameter c: (28) c models the immediate utterance context; it provides the speaker, addressee, time, and place of the current utterance

Interpreting 1st and 2nd sg pronouns Formalizing I and you The formalization that we will adopt was proposed by Kaplan (1977), who introduced a context parameter c: (28) c models the immediate utterance context; it provides the speaker, addressee, time, and place of the current utterance First and second person pronouns access the context parameter to fix their interpretation on an individual. In updated notation: (29) a. I w,g,c = speaker(c) b. you w,g,c = addressee(c)

Interpreting 1st and 2nd sg pronouns Sample derivation DP e You e VP t V e,t laugh e,t w,g,c 2 (NN) = You e VP t laugh e,t w,g,c (FA) = laugh w,g,c (you w,g,c ) = [λx. laugh (x)(w)](addressee(c)) λ = 1 iff laugh (addressee(c))(w)

Summary Summary The interpretation of 3rd sg pronouns is variable and depending on the use draws on the physical or the discourse context. In contrast, 1st and 2nd sg pronouns depend on the immediate utterance context. The variability is captured by having 3rd sg pronouns access an assignment function g that tracks individual that can be referred to. (30) Pronouns Rule (PR): If α is a pronoun, then for any assignment g and index i for which g assigns a value: α i w,g,c = g(i) 1st and 2nd singular pronouns access a context parameter c that encodes the current speaker and addressee.