Introduction to Computational Linguistics

Download Report

Transcript Introduction to Computational Linguistics

Introduction to Linguistics II
Ling 2-121C, group b
Lecture 5
Eleni Miltsakaki
AUTH
Spring 2006
1
Syntax review II
• What are heads and complements? Can you identify
them in the phrases below?
– The man with the telescope
– The destruction of Rome
– A person worthy of praise
– A boy who pitched a perfect game
• What do we mean by selectional restrictions of verbs?
• Are selectional restrictions applicable to verbs only? Can
you give examples of selectional restrictions forced by
other word classes?
2
Syntax review II
• What are phrase structure rules?
• Write 5 basic phrase structure rules for English
• What are the two basic uses of phrase structure rules?
• What is recursion? When does it occur?
• Can you give some examples of recursive structures in
English?
3
Exercise
• The girl with the feather on the ribbon on
the brim
• Expand the above sentence with a
recursive structure and show that in a tree
4
Syntax review II
• What are embedded structures? Give
examples.
• What do we call center embedding? Give
examples.
• What’s special about center embedding?
5
Sentence relatedness
• Transformational rules
– A way to capture the relationship between a
declarative and a question is to allow phrase structure
to generate the structure using special rules:
transformation rules
– Move Aux: take the first aux and move it to the left of
the subject
– The boy is sleeping  Is the boy ___ sleeping?
6
Deep and surface structure
• Deep structure: the basic structure
• Surface structure: the resulting structure
after applying a transformational rule
• The boys is sleeping: Deep
• Move Aux
• Is the boy sleeping? Surface
7
Other transformational rules
• Active passive (aka passivization)
– The cat chased the mouse
– The mouse was chased by the cat
• There-sentences
– There was a man on the roof
– A man was on the roof
• PP-preposing
– The astronomer saw the star with the telescope
– With the telescope, the man saw the star
8
Syntactic dependencies
• Wh-questions
– Who did Helen say the senator wanted to hire ___?
– Who did Helen say the senator wanted the congressional representative
to try to hire ___?
– Who did Helen say the senator wanted the congressional representative
to try to convince the Speaker of the House to get the Vice President to
hire ___?
• Long-distance dependencies created by wh-movement are a
fundamental part of human language.
• They provide evidence that sentences are not just strings of words
but they are supported by rich phrase structure trees. These trees
express the underlying structure of the sentence as well as their
relation to other sentences in the language
9
Universal grammar
• UG provides the basic design for human
language
• Individual languages are variations on the
basic design
10
Parameters
• All languages have phrase structure rules
• All languages have heads and
complements
• In some languages heads come before
complements (English)
• In others, complements come before
heads (Japanese).
11
Wh-islands
• Emily paid a visit to the senator who wants to hire who?
– *Who did Emily pay a visit to the senator that wants to hire ___?
– Miss Marple asked Sherlock whether Poirot had solved the crime
– Who did Miss Marple ask ___ whether Poirot had solved the
crime?
– *Who did Miss Marple ask Sherlock whether ___ had solved the
crime?
– *What did Miss Marple ask Sherlock whether Poirot had solved
___?
12
Universal wh-islands
• Constraints against wh-movement depend
on structure, not length
– John ate bread and cheese (Coordinate).
– John ate bread with cheese (DO).
– *What did John eat bread and?
– What did John eat bologna with?
13
SEMANTICS
14
Intro to semantics
•
What is semantics?
– The study of the linguistic meaning of morphemes, words, phrases, and
sentences.
•
What is lexical semantics?
– The study of the meaning of words and the meaning relations among
them.
•
What is phrasal or sentential semantics?
– The meaning of syntactic units larger than the word
•
What is pragmatic meaning?
– The effect of context on sentence meaning. E.g., “It’s open here”
meaning “Open the window”
15
Lexical semantics
• Learning a language includes learning the
agreed-upon meanings of certain strings of
sounds and how to combine these units into
larger meaningful units.
• The meaning of words is part of linguistic
knowledge/grammar.
• The mental storehouse of the info about words
and morphemes is called the lexicon.
16
Semantic properties
• Example: assassin
– Person
– Who murdered an important person
• Sharing properties
– Female: woman, aunt, mother, tigress, hen
– Human: doctor, bachelor, professor, baby
– Young: child, baby
17
Semantic properties
• Verbs
– Motion: bring, fall, walk, run
– Contact: hit, kick, kiss
– Creation: build, imagine, make
– Sense: hear, see, feel
18
Semantic properties
• Evidence for semantic properties: errors
Bridge of nose
When my gums bled
Bridge of neck
When my tongues bled
He came too late
Mary was young
That’s a horse of
another color
He came too early
Mary was early
That’s a horse of
another race
19
Semantic properties
• Attention!
– Do not confuse semantic properties of words
with nonlinguistic properties.
– E.g., Water is composed of hydrogen and
oxygen.
– But that’s not part of the linguistic meaning of
the word ‘water’.
20
Semantic properties
• One way of representing semantic
properties is by semantic features.
Woman
+female
+human
-young
Father
+male
+human
+parent
Girl
+female
+human
+young
Mare
+female
-human
-young
+equine
21
Homonyms
• Homonyms are different words that are
pronounced the same:
– Tale – tail
– To – two – too
• They may or may not have different spelling.
• Homonyms create ambiguity:
– I’ll meet you by the bank, in front of the automated
teller machine.
– I’ll meet you by the bank. We can go fishing.
22
Polysemy
• When a word has multiple meanings it is
polysemous.
– Bear=to tolerate, to carry, to support
– Bear is also a homonym: bear=wild animal
23
Heteronym
• Two words are heteronyms if they are
spelled the same, but are pronounced
differently.
– Dove – the bird
– Dove – past tense of ‘dive’
– Lead – the verb
– Lead – the metal
24
Homographs
• Homographs are words that are spelled the
same but have different meanings:
– Dove – the bird
– Dove – the past tense of dive
– Bear – the verb
– Bear – the animal
25
Homonyms
Heteronyms
Homographs
Pronounced
identically
Yes
No
Yes/no
Spelled
identically
Yes/no
Yes
Yes
26
Synonyms
• Words that sound different but have the
same (or almost the same) meaning are
synonyms.
– Apathetic, phlegmatic, passive, sluggish,
indifferent
– Sofa, couch
• Degree of similarity depends on number of
semantic properties that two words share
27
Antonyms
• Words that are opposite in meaning are
antonyms
– Beautiful – ugly (=not beautiful)
– Big – small, hot – cold
• Gradable properties
– Tiny, small, medium, large, huge, gigantic
– “not tiny” doesn’t mean “gigantic”.
28
Gradable antonyms
• Marked – unmarked
– How high is the mountain? (not “How low is
the mountain?”)
– “high” is the unmarked member of the pair
“high-low”
29
Relational opposites
• Relational opposites display symmetry in their
meaning: e.g., if X gives Y to Z then Z receives Y
from X
– Give – receive
– Buy – sell
– Teacher – pupil
• Comparative forms of gradable pairs of
adjectives often form relational pairs
– Mary is taller than Sally
– A Mercedes is more expensive than a Smart.
30
• Relationships between certain semantic
features can reveal knowledge about
antonyms:
– A word that is [+married] is [-single]
– A word that is [+single] is [-married]
31
Autoantonyms
• The same word having two antonymic
senses:
– To cleave: to split apart, to cling together
– To dust: to remove sth, to spread sth
32
Formation of antonyms
•
•
•
•
•
UnNonInDisMis-
33