LINGUISTICA GENERALE E COMPUTAZIONALE CONOSCENZA LESSICALE IL LESSICO

LINGUISTICA GENERALE E COMPUTAZIONALE CONOSCENZA LESSICALE: IL LESSICO GENERATIVO

Introduction Lexicon— ideally collection of all words of a language • • Information stored in a lexiconØ Phonetic information § pronunciation Ø Semantic information § meaning Ø Morphological information § transitivity and intransitivity (verbs) , count vs. mass (noun) 2

Lexicon (contd…) Example of “eat” in the Oxford Advanced Learner’s Dictionary eat /i: t/ v (pt ate /et/; pp eaten /i: tn/): 1. sth (up) to food into the mouth, chew and swallow it: he was too i eat Pronunciation Morphological information Meaning Lexical entry 3

Mental Lexicon • Mental Lexicon: information stored in the mind of a native speaker • Native speakers store information Ø Phonetic information § pronunciation Ø Semantic information § meaning Ø Morphological information § transitivity vs. intransitivity (verbs), count vs. mass (noun) • Additional information § use of a word in a new context, syntactic environment of a word, word-formation rules 4

Example of Mental Lexicon Example of eat in a native speaker’s mind • Pronunciation: long /i: / is used in eat • Grammatical information: past tense is ate /et/ • Word-formation rules: /-s/ is the third person singular present tense marker as in he eats • Meaning: 1. Take in solid food: she ate a banana 2. Take a meal: we did not eat until 10 P. M. 3. Worry or cause anxiety in a persistent way: what’s eating you up. • Syntactic Information: eat needs an agent to perform the action. the agent role is obligatory. 5

Lexicon in Computational Linguistics Lexicon meant for Natural Language Processing (NLP) must have the following properties: • Morphological information Ø Parts of speech information Ø Rules should be there to deal with both regular and irregular forms e. g ate (past tense of eat) men (plural of man) • Semantic information Ø • Can handle lexical ambiguity Syntactic information Ø Action verbs will always have an agent 6

Polysemy and the Logical Problem of Polysemy • An individual word can have indefinite number of subtle meaning difference • Natural Languages are highly polysemous • This creates ambiguity • Weinreich distinguishes between two types of ambiguity Ø Contrastive ambiguity Ø Complementary polysemy 7

Polysemy and the Logical Problem of Polysemy (contd…) Contrastive Ambiguity • • A lexical item carries two distinct unrelated meanings This is a case of homonymy Ø words spelled or pronounced in the same way but have different meanings Example: Ø bank a financial institution Ø bank place beside a body of water. 8

Complementary polysemy (1)a. Mary doesn’t believe the book. b. John sold his books to Mary. (2)a. Eno the cat is sitting on yesterday’s newspaper. b. Yesterday’s newspaper really got me upset. (3)a. Mary is in Harvard Square looking for the Bach sonatas. b. We won’t get to the concert until after the Bach sonatas. (4)a. I have my lunch in the backpack. b. Your lunch was no longer today than it was yesterday. (5)a. The phone rang during my appointment. b. My next appointment is John.

Sense Enumeration Lexicon (SEL) • Word. Net and similar resources are examples of SENSE ENUMERATION LEXICA • Direct approach to handle polysemy is to allow the lexicon to have multiple listing of words, each annotated with a separate meaning or lexical sense. • Widely accepted in both computational and theoretical linguistics. 11

Sense Enumeration Lexicon (SEL) • Example of Contrastive Senses bank 1 CAT= count-noun GENUS= financial-institution bank 2 CAT= count-noun GENUS= shore 12

Nominal polysemy in sense enumeration lexica • Newspaper 1 CAT= count-noun GENUS= artefact Newspaper 2 CAT= count-noun GENUS= information 13

Sense Enumeration Lexicon (SEL) • Possible Modification of Complementary Polysemy in SEL newspaper sense 1 CAT= count-noun GENUS= artefact sense 2 CAT= count-noun GENUS= information 14

Verbal polysemy in sense enumeration lexica Syntactic polysemy deals with polivalency (I), object deletion (II) and the general properties of argument expression (III). (I) a. Mary began to read the novel. b. Mary began reading the novel. c. Mary began the novel. (II) a. Mary ate (her meal) quickly. b. Mary devoured *(her meal) quickly. (III) a. John carved a doll (out of the wood). b. John carved the wood (into a doll).

GENERATIVE LEXICON THEORY • (Pustejovsky, 1991, 1995) • Claim: the concepts associated with a word in a context are GENERATED by a process starting from lexical entries structured into QUALIA STRUCTUREs and involving GENERATIVE DEVICES such as TYPE COERCION and CO-COMPOSITION

Generative lexicon theory: lexical entries A lexical entry in the generative lexicon consists of the following elements at least: • Argument Structure Ø True Arguments Ø Default Arguments Ø Shadow Arguments Ø True Adjuncts • Event Structure • Qualia Structure Ø Formal Ø Constitutive Ø Telic Ø Agentive 19

Argument Structure • True Arguments: syntactically realized parameters of the lexical item John arrived late • Default Arguments: logically present in the expressions but are not necessarily expressed syntactically. John built the house out of bricks • True Adjuncts: Ø modify the logical expression Ø part of the situational interpretation She drove down to New York on Tuesday. 20

Argument Structure (contd…) • Shadow Arguments: semantically incorporated in the lexical item and are expressed by discourse specification and contextual factors Mary buttered her toast Ø hidden argument is the material being spread on the toast Ø these are not optional arguments but expressible only under specific conditions Ø refer to the semantic content that is not necessarily expressed in syntax Example: Mary buttered her toast with margarine 21

SELECTION 1. a. The man fell/died. b. The rock fell/!died. 2. a. John forced/!convinced the door to open. b. John forced/convinced the guests to leave. 3. a. John poured milk into /!on his coffee. b. John poured milk into/on the bowl.

Integrating Selection into Grammars

CONSEQUENCE OF SELECTION: ONTOLOGICAL ASSUMPTIONS

BUT … • Unlike in other lexical theories, in GLT types can be modified via TYPE COERCION (see below)

Event Structure • event type of a lexical item and a phrase • events can be sub-classified into at least three sorts: State, Process and Transition Event Structure of build as found in the following expressions They are building a new house The house was built by John build EVENTSTR= E 1= process E 2= state 26

QUALIA ROLES Word meaning is structured on the basis of four generative factors, called qualia roles, that capture how humans understand objects and relations in the world and provide the minimal explanation for the linguistic behavior of lexical items. FORMAL: the basic category that distinguishes an object within a larger domain CONSTITUTIVE: the relation between an object and its constituent parts TELIC: the object’s purpose and function AGENTIVE: factors involved in the object’s origin or ‘coming to being’

Qualia structures and argument polysemy Qualia Structure for novel Qualia const = narrative formal = book telic = reading agent = writing 28

A generative device: type coercion • Type Coercion Ø a lexical item or phrase is coerced to a semantic interpretation by a governing item in the phrase, without changing its syntactic type Mary began to read the novel Mary began reading the novel Mary began the novel • • Function Application with Coercion Ø different complement type of the verb Ø different interpretations of the verb that arise for the different complements 29

Other generative devices • Selective Binding Ø a lexical item or a phrase operates specifically on the substructure of a phrase, without changing the overall type in the composition a good knife: a knife that cuts well • Co-composition Ø multiple elements within a phrase behave as functors, generating new nonlexicalized senses for the words in composition John baked the potato John baked the cake 30

NOMINAL POLYSEMY A dot object is a deeper structure relating the apparently contradictory senses of the word. For each sense pair there is a relation that ‘connects’ the senses in a well-defined way. The dot object is characterized as: - a Cartesian type product of n types (the product τ1 x τ2, of types τ1 and τ2, each denoting sets, is the ordered pair , where t 1 ε τ1 , t 2 ε τ2) - with some additional constraints: there exists a relation R between the elements of τ1 and τ2 , namely, R(t 1 , t 2). This relation must be seen as part of the definition of the semantics for the dot object.

Type combinations included in the broad range of complex types encountered in natural language: a. b. c. d. e. phys_obj info : e. g. , book, record event : e. g. , construction, examination event question : e. g. , exam event food : e. g. , lunch, dinner event human : e. g. , appointment For each of these type products, there is a unique relation, Ri, that structures the types. For example, nouns such as book or record, are structured by a containment relation R (container-like concepts). This containment relation -hold(x, y)- must be encoded directly into the semantics of the concept as the FORMAL quale value.

The lexical structure for newspaper as a dot object is represented as follows: newspaper ARGSTR = ARG 1 = y: information ARG 2 = x: phys_obj QUALIA = information phys_obj FORM = hold(x, y) TELIC = read(e, w, x y) AGENT = write(e, v, x y) This translates to the following logical expression: λx y e v[newspaper (x: physobj y : info) hold(x, y) λwλe [read (e, w, x y) [write(e; v, x y)]]

MORE FORMAL DETAILS

Three Ranks of Type Entities Events

System of Generating Types

Qualia are incorporated into Type Itself

Qualia as Types

Functional Selection

Functional Type Coercion

Co-composition

Coercion in Function Composition

Selection and Coercion

Type Specification

GLT AND LEXICAL RESOURCES

Generative lexicon vs. Word. Net • Formal role is similar to the hypernymy relation • Constitutive role is similar to the meronymy relation • Nothing in Word. Net like • the functionality link • Event structure • Exists in some Word. Nets, e. g. , Hindi Word. Net 50

LEXICAL RESOURCES BASED ON GLT • SIMPLE • LKB

SIMPLE • Lessico creato all’Universita’ di Pisa

some semantic types for abstract & concrete entities TELIC AGENTIVE TOP CONSTITUTIVE . . . Representation . . . Concrete_entity ENTITY • Living_entity Event. . . Property Abstract_entity • Sign • Human • Quality • Language • Animal • Quantity • Information • Vegetal_entity • . . . • Artifact • Convention • Cognitive_fact • Physical_prop • . . . • Psychol_prop Artifact • . . . • Susbstance • Furniture • Instrument • Location • Food • Clothing • Material • Artwork Artifactual_material

some semantic types for events EVENT Phenomenon Aspectual. . State Act . . . Relational_state. . . Non_relational_act Relational_act Cause_change Psych_event Change. . . Relational_change Move Change_possession Cause_act Speech_act. . . Creation. . . Acquire_knowledge Natural_transition Change_location

some semantic types for adjectives TOP Intensional Temporal Extensional Psychological_prop Modal Emotive Relational_prop Social_prop Emphasizer Manner Object_related Physical_prop Intensifying_prop Temporal_prop

E Formal x t isa antonym_comp e antonym_grad mult_opposition n d. Q E e u x d a t e l n i S d a t e r d u c r t o u l r e e s Constitutive made_of is_a_follower_of has_as_member is_a_member_of has_as_part instrument kinship is_a_part_of resulting_state relates uses C O N S T I T U T I V E causes concerns affects constitutive_activity P contains R has_as_colour has_as_effect O has_as_property P measured_by E measures R produces produced_by T property_of Y quantifies related_to successor_of precedes typical_of contains feeling is_in lives_in LOCATION typical_location Agentive result_of agentive_prog agentive_cause agentive_experience caused_by source created_by derived_from Telic A G E N T I V E ARTIFACTUAL AGENTIVE used_for used_as used_by used_against INSTRUMENTAL indirect_telic purpose is_the_activity_of is_the_ability_of is_the_habit_of object_of_activity TELIC ACTIVITY DIRECT TELIC

Orthogonal dimensions of meaning c ro le Agentive role used_for Teli y ed_b is_made_of instrument crea t Con stitu tive role is_a Formal role

Orthogonal dimensions of meaning Agentive role playing le c ro used_for Teli mak e by violin ed_ wood strings ical _ins mus has_as_part is_made_of crea t Con stitu tive is_a role trum ent Formal role

meaning dimensions expressed by Qualia relations botte barrel recipiente di legno traditional dictionary definition fatto Formal: isa Agentive: created_by Constitutive: made_of di doghe arcuate tenute unite da cerchi di ferro che serve per la conservazione e il trasporto di liquidi, specialmente vino Constitutive: contains Telic: Used_for

REFERENCES • Pustejovsky, J. (1995). The generative Lexicon. Cambridge, MA: MIT Press.

ACKNOWLEDGMENTS • Slides borrowed from – Debasri Chakrabarti – James Pustejovsky – Nilda Ruimy