Embodied Compositional Semantics Ellen Dodge

Slides:



Advertisements
Similar presentations
CSCTR Session 8 Dana Retová. group at UC Berkeley & Uni of Hawaii Nancy Chang Benjamin Bergen Jerome Feldman, … General assumption Semantic relations.
Advertisements

 Christel Kemke 2007/08 COMP 4060 Natural Language Processing Feature Structures and Unification.
Lexical Functional Grammar : Grammar Formalisms Spring Term 2004.
1 Natural Language Processing Lecture 7 Unification Grammars Reading: James Allen NLU (Chapter 4)
English 306A; Harris 1 Sentences Homer preens. Homer loves his blanket. Homer eats. Homer eats crap. The fat bald guy eats crap. The fat bald guy eats.
Syntax-Semantics Mapping Rajat Kumar Mohanty CFILT.
What does language do? “Harry walked to the cafe.” “Harry walked into the cafe.” A sentence can evoke an imagined scene and resulting inferences : CAFE.
Embodied Construction Grammar ECG (Formalizing Cognitive Linguistics) 1.Community Grammar and Core Concepts 2.Deep Grammatical Analysis 3.Computational.
Language and Cognition Colombo, June 2011 Day 2 Introduction to Linguistic Theory, Part 4.
Lectures I. Overview 2. Simulation Semantics 3. ECG and Best-fit Analysis 4. Compositionality 5. Simulation, Counterfactuals, and Inference Constructions.
Statistical NLP: Lecture 3
PSY 369: Psycholinguistics Some basic linguistic theory part2.
Steven Schoonover.  What is VerbNet?  Levin Classification  In-depth look at VerbNet  Evolution of VerbNet  What is FrameNet?  Applications.
The Case for Case Reopened ‘Agents and Agency Revisited’
1 Pertemuan 23 Syntatic Processing Matakuliah: T0264/Intelijensia Semu Tahun: 2005 Versi: 1/0.
NLP and Speech 2004 Feature Structures Feature Structures and Unification.
Language, Mind, and Brain by Ewa Dabrowska Chapter 10: The cognitive enterprise.
The Neural Basis of Thought and Language Final Review Session.
Using Interfaces to Analyze Compositionality Haiyang Zheng and Rachel Zhou EE290N Class Project Presentation Dec. 10, 2004.
CS 182 Sections slides created by: Eva Mok modified by jgm April 26, 2006.
CS 182 Sections Eva Mok April 21, 2004.
Cognitive Linguistics Croft & Cruse 10 An overview of construction grammars (part 1, through )
Understanding Sentences. Two steps back: What is linguistic knowledge? Phonological Syntactical Morphological Lexical Semantic.
The syntax of language How do we form sentences? Processing syntax. Language and the brain.
Software Testing and QA Theory and Practice (Chapter 4: Control Flow Testing) © Naik & Tripathy 1 Software Testing and Quality Assurance Theory and Practice.
3.2 friction Friction is a force that opposes motion.
Phonetics, Phonology, Morphology and Syntax
Jerome Feldman Simulation Semantics, Embodied Construction Grammar, and the Language of Actions and Events Jerome Feldman
NLU: Frames Frame KR is a good way to represent common sense –can define stereotypical aspects of some domain we are interested in analyzing –sentences.
Comments to Karin Kipper Christiane Fellbaum Princeton University and Berlin-Brandenburgische Akademie der Wissenschaften.
NTL – Converging Constraints Basic concepts and words derive their meaning from embodied experience. Abstract and theoretical concepts derive their meaning.
Syntax Lecture 8: Verb Types 1. Introduction We have seen: – The subject starts off close to the verb, but moves to specifier of IP – The verb starts.
FACULTY OF ENGLISH LANGUAGE AND LITERATURE G. TOGIA SECTION ΠΗ-Ω 10/14/2009 Introduction to linguistics II 1.
Copyright © 2013 Curt Hill The Zachman Framework What is it all about?
Simulation-based language understanding “Harry walked to the cafe.” SchemaTrajectorGoal walkHarrycafe Analysis Process Simulation Specification Utterance.
What Makes an Essay an Essay. Essay is defined as a short piece of composition written from a writer’s point of view that is most commonly linked to an.
Overview of MOT Knowledge representation system : Basic Modeling Editor LexiconGrammarSemantics Pragmatics MOT Editor.
Introduction to Embodied Construction Grammar March 4, 2003 Ben Bergen
Linguistic Essentials
The Minimalist Program
Embodiment & Compositionality Two fundamental, but unreconciled, aspects of human language: embodiment and compositionality Embodiment – the realization.
Fact Extraction Ontology Ontological- Semantic Analysis Text Meaning Representation (TMR) Fact Repository (FR) Text Sources Lexicons Grammars Static Knowledge.
11 Project, Part 3. Outline Basics of supervised learning using Naïve Bayes (using a simpler example) Features for the project 2.
NLP. Introduction to NLP (U)nderstanding and (G)eneration Language Computer (U) Language (G)
NLP. Introduction to NLP Last week, Min broke the window with a hammer. The window was broken with a hammer by Min last week With a hammer, Min broke.
An Introduction to Semantic Parts of Speech Rajat Kumar Mohanty rkm[AT]cse[DOT]iitb[DOT]ac[DOT]in Centre for Indian Language Technology Department of Computer.
Principles and Parameters (II) Rajat Kumar Mohanty Department of Computer Science and Engineering Indian Institute of Technology Bombay.
September 26, : Grammars and Lexicons Lori Levin.
MENTAL GRAMMAR Language and mind. First half of 20 th cent. – What the main goal of linguistics should be? Behaviorism – Bloomfield: goal of linguistics.
Lec. 10.  In this section we explain which constituents of a sentence are minimally required, and why. We first provide an informal discussion and then.
Getting From the Utterance to the SemSpec Johno Bryant Need a grammar formalism –Embodied Construction Grammar (Bergen & Chang 2002) Need new models for.
The Neural Basis of Thought and Language Final Review Session.
The Neural Basis of Thought and Language Week 14.
CS 182 Sections Leon Barrett with slides inspired by Eva Mok and Joe Makin April 18, 2007.
The Neural Basis of Thought and Language Week 14.
A Compositional Constructional Analysis of ‘Hitting’ Verb Argument Realization Patterns and Their Meanings Ellen K. Dodge International Computer Science.
CS182 Leon Barrett with slides inspired by Eva Mok and Joe Makin April 25, 2008.
CS 182 Sections slide credit to Eva Mok and Joe Makin Updated by Leon Barrett April 25, 2007.
Grammar Grammar analysis.
กำหนดการเรียนการสอนครั้งที่ ๙
Lecture – VIII Monojit Choudhury RS, CSE, IIT Kharagpur
Passive and Active Voice
Statistical NLP: Lecture 3
Lexical Functional Grammar
Strong Slot-and-Filler Structures
CSC 594 Topics in AI – Applied Natural Language Processing
Cognitive Language Comprehension in Rosie
Strong Slot-and-Filler Structures
Linguistic Essentials
Structure of a Lexicon Debasri Chakrabarti 13-May-19.
Presentation transcript:

Embodied Compositional Semantics Ellen Dodge

Questions What is the nature of compositionality in the Neural Theory of Language? How can it be best represented using Embodied Construction Grammar?

Examples He bit the apple He was bitten (by a toddler) He bit into the apple His white teeth bit into the apple. He shattered the window The window was shattered The window shattered

Outline Compositionality Neural Theory of Language and ECG –Assumptions –Overview Examples: –Representation of constructions and meaning –Simulation Concluding Remarks

Compositionality Put the parts together to create the meaning of the whole.

Compositionality Put the parts together to create the meaning of the whole. Questions: –what is the nature of the parts? –How and why do they combine with one another? –What meaning is associated with this composition?

Short answers Parts = constructions, schemas Combination = binding, unification Meaning of the whole = simulation of unified parts

Constructions Construction Grammar Constructions are form-meaning pairings A given utterance instantiates many different constructions Embodied Construction Grammar Construction meaning is represented using schemas Meaning is embodied

Key assumptions of NTL Language understanding is simulation Simulation involves activation of neural structures

Comments Language understanding Understanding process is dynamic “Redundancy” is okay

Conceptual structure Embodied Schematic (Potentially) language-independent Highly interconnected

Simulation parameters Constructions unify to create semantic specification that supports a simulation Two types of simulation parameters for event descriptions: –Event content –Event construal

Putting the parts together Bindings Unification

“Pre-existing” structure Cxn schema Cxn schema

Unification Cxn schema Cxn schema

Summary Parts = constructions, schemas Combination = binding, unification Meaning of the whole = simulation of the combined parts

First example He bit the apple.

schema MotorControl subcase of Process roles Actor ↔ Protagonist Effector Effort Routine constraints Actor ← animate Schemas

schema ForceApplication subcase of MotorControl evokes ForceTransfer as FT roles Actor ↔ FT.Supplier ↔ Protagonist Acted Upon↔ FT.Recipient Effector Routine Effort ↔ FT.Force.amount schema ForceTransfer evokes Conact as C roles Supplier ↔ C.entity1 Recipient ↔ C.entity2 Force schema MotorControl subcase of Process roles Actor ↔ Protagonist Effector Effort Routine constraints Actor ← animate schema Contact subcase of SpatialRelation roles Entity1 : entity Entity2 : entity

Schema networks MotorControl Motion SPG Effector Motion Effector MotionPath ForceTransfer ForceApplication Contact SpatiallyDirectedAction CauseEffect Contact Agentive Impact SelfMotion Path MotionPath

Construction BITE1 subcase of Verb form: bite meaning: ForceApplication constraints: Effector ← teeth Routine ← bite // close mouth Verb Constructions schema ForceApplication subcase of MotorControl evokes ForceTransfer as FT roles Actor ↔ FT.Supplier ↔ Protagonist Acted Upon ↔ FT.Recipient Effector Routine Effort ↔ FT.Force.amount

Verb Constructions schema ForceApplication subcase of MotorControl schema Agentive Impact subcase of ForceApplication cxn BITE meaning: ForceApplication schema MotorControl cxn GRASP meaning: ForceApplication cxn PUSH meaning: ForceApplication cxn SLAP meaning: AgentiveImpact cxn KICK meaning: AgentiveImpact cxn HIT meaning: AgentiveImpact

Argument Structure Construction construction ActiveTransitiveAction2 subcase of VP constituents: V : verb NP: NP form constraints: V F before NP F meaning: CauseEffect evokes; EventDescriptor as ED; ForceApplication as FA constraints: {Self m ↔ ED.EventType} {V m ↔ ED.ProfiledProcess} Causer ↔ ED.ProfiledParticipant FA ↔ V m Causer ↔ FA.Actor Affected ↔ FA.ActedUpon Affected ↔ NP m

Argument Structure Construction construction ActiveTransitiveAction2 subcase of VP constituents: V : verb NP: NP form constraints: V F before NP F meaning: CauseEffect evokes; EventDescriptor as ED; ForceApplication as FA constraints: {Self m ↔ ED.EventType} {V m ↔ ED.ProfiledProcess} Causer ↔ ED.ProfiledParticipant FA ↔ V m Causer ↔ FA.Actor Affected ↔ FA.ActedUpon Affected ↔ NP m

CauseEffect schema schema CauseEffect subcase of ForceApplication; Process roles Causer ↔ Actor Affected ↔ ActedUpon ↔ Process.Protagonist Instrument ↔ Effector

MotorControl Motion SPG Effector Motion Effector MotionPath ForceTransfer ForceApplication Contact SpatiallyDirectedAction CauseEffect Contact SelfMotion Path MotionPath Agentive Impact Process Schema Network

Argument Structure Construction construction ActiveTransitiveAction2 subcase of VP constituents: V : verb NP: NP form constraints: V F before NP F meaning: CauseEffect evokes: EventDescriptor as ED; ForceApplication as FA constraints: {Self m ↔ ED.EventType} {V m ↔ ED.ProfiledProcess} Causer ↔ ED.ProfiledParticipant FA ↔ V m Causer ↔ FA.Actor Affected ↔ FA.ActedUpon Affected ↔ NP m

MotorControl Motion SPG Effector Motion Effector MotionPath ForceTransfer ForceApplication Contact SpatiallyDirectedAction CauseEffect Contact SelfMotion Path MotionPath Agentive Impact Process Schema Network

Important points  Compositionality does not require that each component contain different information.  Shared semantic structure is not viewed as an undesirable redundancy

Argument Structure Construction construction ActiveTransitiveAction2 subcase of VP constituents: V : verb NP: NP form constraints: V F before NP F meaning: CauseEffect evokes; EventDescriptor as ED ; ForceApplication as FA constraints: {Self m ↔ ED.EventType} {V m ↔ ED.ProfiledProcess} Causer ↔ ED.ProfiledParticipant FA ↔ V m Causer ↔ FA.Actor Affected ↔ FA.ActedUpon Affected ↔ NP m

schema EventDescriptor roles EventType: Process ProfiledProcess: Process ProfiledParticipant: Entity ProfiledState(s): State SpatialSetting TemporalSetting Event Descriptor schema

Argument Structure Construction Construction ActiveTransitiveAction2 subcase of VP constituents: V : verb NP: NP form constraints: V F before NP F meaning: CauseEffect evokes; EventDescriptor as ED ; ForceApplication as FA constraints: {Self m ↔ ED.EventType} {V m ↔ ED.ProfiledProcess} Causer ↔ ED.ProfiledParticipant FA ↔ V m Causer ↔ FA.Actor Affected ↔ FA.ActedUpon Affected ↔ NP m

construction NPVP1 constituents: Subj: NP VP : VP form Constraints Subj f before VP f meaning: EventDescriptor ProfiledParticipant ↔ Subj m Bindings with other cxns construction ActiveTransitiveAction2 subcase of VP constituents: V ; NP form: V F before NP F meaning: CauseEffect evokes; EventDescriptor as ED constraints: {Self m ↔ ED.EventType} {V m ↔ ED.ProfiledProcess} Causer ↔ ED.ProfiledParticipant Affected ↔ NP m

Construction NPVP1 constituents: Subj: NP VP : VP form constraints Subj f before VP f meaning: EventDescriptor ProfiledParticipant ↔ Subj m Bindings with other cxns construction ActiveTransitiveAction2 subcase of VP constituents: V ; NP form: V F before NP F meaning: CauseEffect evokes; EventDescriptor as ED constraints: {Self m ↔ ED.EventType} {V m ↔ ED.ProfiledProcess} Causer ↔ ED.ProfiledParticipant Affected ↔ NP m schema EventDescriptor roles EventType ProfiledProcess ProfiledParticipant ProfiledState(s) SpatialSetting TemporalSetting

Bindings with other cxns schema EventDescriptor roles EventType ProfiledProcess ProfiledParticipant ProfiledState(s) SpatialSetting TemporalSetting construction NPVP1 constituents: Subj: NP VP : VP form Constraints Subj f before VP f meaning: EventDescriptor ProfiledParticipant ↔ Subj m construction ActiveTransitiveAction2 subcase of VP constituents: V ; NP form: V F before NP F meaning: CauseEffect evokes; EventDescriptor as ED constraints: {Self m ↔ ED.EventType} {V m ↔ ED.ProfiledProcess} Causer ↔ ED.ProfiledParticipant Affected ↔ NP m

Unification CauseEffect causer affected ForceApplication actor actedupon EventDescriptor EventType ProfiledProcess ProfiledParticipant BITE TransitiveAction2 HE NP1 NPVP1 THEAPPLE NP2 ReferentDescriptor ReferentDescriptor MeaningConstructions

Unification CauseEffect causer affected ForceApplication actor actedupon EventDescriptor EventType ProfiledProcess ProfiledParticipant BITE TransitiveAction2 HE NP1 NPVP1 THEAPPLE NP2 ReferentDescriptor ReferentDescriptor resolved referent MeaningConstructions

Unification CauseEffect causer affected ForceApplication actor actedupon EventDescriptor eventtype ProfiledProcess ProfiledParticipant BITE TransitiveAction2 Verb HE NP1 NPVP1 THEAPPLE NP2 ReferentDescriptor ReferentDescriptor resolved referent MeaningConstructions

Unification CauseEffect causer affected ForceApplication actor actedupon EventDescriptor eventtype ProfiledProcess ProfiledParticipant BITE TransitiveAction2 HE NP1 NPVP1 subj THEAPPLE NP2 ReferentDescriptor ReferentDescriptor MeaningConstructions

Unification CauseEffect causer affected ForceApplication actor actedupon EventDescriptor eventtype ProfiledProcess ProfiledParticipant BITE TransitiveAction2 NP HE NP1 NPVP1 THEAPPLE NP2 ReferentDescriptor ReferentDescriptor MeaningConstructions

Semantic Specification He bit the apple EventDescriptor eventtype ProfiledProcess ProfiledParticipant CauseEffect causer affected ForceApplication actor actedupon routine  bite effector  teeth RD55 category Person Apple RD27 category

Process Simulation - He bit the apple CauseEffect ForceApplication Protagonist = Causer ↔ Actor Protagonist = Affected ↔ ActedUpon

Process Simulation - He bit the apple CauseEffect ForceApplication Protagonist = Causer ↔ Actor Protagonist = Affected ↔ ActedUpon

Passive voice He was bitten (by a toddler)

Argument Structure Construction He was bitten (by a toddler) construction PassiveTransitiveAction2 subcase of VP constituents: V : PassiveVerb (PP: agentivePP) form constraints: V F before PP F meaning: CauseEffectAction evokes; EventDescriptor as ED; ForceApplication as FA constraints: {Self m ↔ ED.EventType} {V m ↔ ED.ProfiledProcess} Affected ↔ ED.ProfiledParticipant FA ↔ V m Causer ↔ FA.Actor Affected ↔ FA.ActedUpon Causer ↔ PP.NP m

Semantic Specification He was bitten (by a toddler) EventDescriptor eventtype ProfiledProcess ProfiledParticipant CauseEffect causer affected ForceApplication actor actedupon routine  bite effector  teeth RD48 category Person Person RD27 category

Effect = Process Simulation - He was bitten (by a toddler) CauseEffect Action = Bite Protagonist = Causer ↔ Actor Protagonist = Affected ↔ ActedUpon

Variations on a theme He shattered the window The window was shattered The window shattered

Construction SHATTER1 subcase of Verb form: shatter meaning: StateChange constraints: Initial :: Undergoer.state ← whole Final :: Undergoer.state ← shards Verb Construction -- shatter schema StateChange subcase of Process roles Undergoer ↔ Protagonist

Argument Structure Construction He shattered the window construction ActiveTransitiveAction3 subcase of VP constituents: V : verb NP: NP form constraints: V F before NP F meaning: CauseEffect evokes: EventDescriptor as ED; StateChange as SC constraints: {Self m ↔ ED.EventType} {V m ↔ ED.ProfiledProcess} Causer ↔ ED.ProfiledParticipant SC ↔ V m Affected ↔ SC.Undergoer Affected ↔ NP m

Semantic Specification He shattered the window EventDescriptor eventtype ProfiledProcess ProfiledParticipant CauseEffect causer affected StateChange Undergoer state  “wholeness” RD189 category Person Window RD27 category

Process Simulation - He shattered the window CauseEffect Action Protagonist = Causer Protagonist = Affected ↔ Undergoer

Argument Structure Construction The window was shattered construction PassiveTransitiveAction3 subcase of VP constituents: V : PassiveVerb (PP: agentivePP) form constraints: V F before NP F meaning: CauseEffect evokes: EventDescriptor as ED; StateChange as SC constraints: {Self m ↔ ED.EventType} {V m ↔ ED.ProfiledProcess} Affected ↔ ED.ProfiledParticipant SC ↔ V m Affected ↔ SC.Undergoer Causer ↔ PP.NP m

Semantic Specification The window was shattered EventDescriptor eventtype ProfiledProcess ProfiledParticipant CauseEffect causer affected StateChange Undergoer state  “wholeness” RD175 category Window

Process Simulation - The window was shattered CauseEffect Action Protagonist = Causer Protagonist = Affected ↔ Undergoer

Argument Structure Construction The window shattered construction ActiveIntransitiveAction1 subcase of VP constituents: V : verb form meaning: Process evokes: EventDescriptor as ED; StateChange as SC constraints: {Self m ↔ ED.EventType} {V m ↔ ED.ProfiledProcess} Protagonist ↔ ED.ProfiledParticipant SC ↔ V m Protagonist ↔ SC.Undergoer

Semantic Specification The window shattered EventDescriptor eventtype ProfiledProcess ProfiledParticipant Process protagonist StateChange Undergoer state  “wholeness” RD177 category Window

Process Simulation - The window shattered Process Protagonist = Undergoer

Some more variations on a theme He bit the apple He bit into the apple His white teeth bit into the apple.

Argument Structure Construction He bit into the apple construction ActiveEffectorMotionPath2 subcase of VP constituents: V : verb PP: Spatial-PP form constraints: V F before PP F meaning: EffectorMotionPath evokes; EventDescriptor as ED; ForceApplication as FA constraints: {Self m ↔ ED.EventType} {V m ↔ ED.ProfiledProcess} Actor ↔ ED.ProfiledParticipant FA ↔ V m Actor ↔ FA.Actor Effector ↔ FA.Effector // INI Target ↔ FA.ActedUpon SPG ↔ PP m Target ↔ PP m.Prep.LM

Schema schema EffectorMotionPath subcase of EffectorMotion subcase of SPG // or evokes SPG roles Actor ↔ MotorControl.protagonist Effector ↔ SPG.Tr ↔ M.Mover ↔ Motion.protagonist Target ↔ SPG.Lm

MotorControl Motion SPG Effector Motion Effector MotionPath ForceTransfer ForceApplication Contact SpatiallyDirectedAction CauseEffect Contact SelfMotion Path MotionPath Agentive Impact Process Schema Network

Argument Structure Construction He bit into the apple construction ActiveEffectorMotionPath2 subcase of VP constituents: V : verb PP: Spatial-PP form constraints: V F before PP F meaning: EffectorMotionPath evokes: EventDescriptor as ED; ForceApplication as FA constraints: {Self m ↔ ED.EventType} {V m ↔ ED.ProfiledProcess} Actor ↔ ED.ProfiledParticipant FA ↔ V m Actor ↔ FA.Actor Effector ↔ FA.Effector // INI Target ↔ FA.ActedUpon SPG ↔ PP m Target ↔ PP m.Prep.LM

EffectorMotionPath Action SourcePathGoal Effector Motion Protagonist = Actor Protagonist = Effector

Argument Structure Construction He bit into the apple construction ActiveEffectorMotionPath2 subcase of VP constituents: V : verb PP: Spatial-PP form constraints: V F before PP F meaning: EffectorMotionPath evokes; EventDescriptor as ED; ForceApplication as FA constraints: {Self m ↔ ED.EventType} {V m ↔ ED.ProfiledProcess} Actor ↔ ED.ProfiledParticipant FA ↔ V m Actor ↔ FA.Actor Effector ↔ FA.Effector // INI Target ↔ FA.ActedUpon SPG ↔ PP m Target ↔ PP m.Prep.LM

Simulation: He bit into the apple Action SourcePathGoal Effector Motion Protagonist = Actor Protagonist = Effector

Argument Structure Construction His white teeth bit into the apple construction ActiveEffectorMotionPath3 subcase of VP constituents: V : verb PP: Spatial-PP form constraints: V F before PP F meaning: EffectorMotionPath evokes; EventDescriptor as ED; ForceApplication as FA constraints: {Self m ↔ ED.EventType} {V m ↔ ED.ProfiledProcess} Effector ↔ ED.ProfiledParticipant FA ↔ V m Actor ↔ FA.Actor // INI Effector ↔ FA.Effector Target ↔ FA.ActedUpon SPG ↔ PP m Target ↔ PP m.Prep.LM

Simulation: His white teeth bit into the apple Action SourcePathGoal Effector Motion Protagonist = Actor Protagonist = Effector

Non-agentive biting He landed on his feet, hitting the narrow pavement outside the yard with such jarring impact that his teeth bit into the edge of his tongue. [BNC] The studs bit into Trent's hand. [BNC] His chest burned savagely as the ropes bit into his skin. [BNC]

MotorControl Motion SPG Effector Motion Effector MotionPath ForceTransfer ForceApplication Contact SpatiallyDirectedAction CauseEffect Contact SelfMotion Path MotionPath Agentive Impact Process Schema Network

Simulation: His teeth bit his tongue SourcePathGoal Motion Protagonist = Mover

Summary Small set of constructions and schemas Composed in different ways Unification produces specification of parameters of simulation Sentence understanding is simulation Different meanings = different simulations