Ch.3 Linguistic Essentials

1
Ch.3 Linguistic Essentials

• ????? ????? ???
• 2000. 1. 13
• ???

2
Contents

• Parts of Speech and Morphology
• Nouns and pronouns
• Determiners and adjectives
• Verbs
• Other parts of speech
• Phrase Structure
• Phrase structure grammars
• Dependency
• X theory
• Phrase structure ambiguity
• Semantics and Pragmatics

3
Parts of Speech and Morphology (1/4)

• Parts of Speech(POS)
• Classes of the words which show similar syntactic
  behavior, and often a typical semantic type
• Syntactic category, Grammatical category
• 3 important POS noun, verb, adjective
• Many words have multiple parts of speech
• Ex) candy noun, verb
• Substitution test
• The most basic test for words belonging to the
  same class
• (3.1)
• POS Tags
• Well-established sets of abbreviations for naming
  POS
• Ex) adjectives JJ, (in the Brown corpus)

4
Parts of Speech and Morphology (2/4)

• Open (lexical) vs Closed (functional) category

5
Parts of Speech and Morphology (3/4)

• Morphological process
• Ex) the formation of the plural form from the
  singular form of the noun (dog -gt dogs)
• Morphology is important in NLP because language
  is productive
• Ex) understanding the syntactic and semantic
  properties of unknown(new) words
• Highly inflecting languages like Finnish
• In English regular verbs (4 distinct form),
  irregular verbs (at most 8 forms)
• In Finnish more than 10,000 forms
• To handle morphology is absolutely essential
• The major types of morphological process
• Inflection, Derivation, Compounding

6
Parts of Speech and Morphology (4/4)

• Inflection
• The systematic modifications of a root form by
  means of prefixes and suffixes to indicate
  grammatical distinctions
• Does not change word class or meaning, but varies
  features such as tense, number, and plurality
• Derivation
• less systematic
• Ex) oldly, difficultly
• more radical change of POS
• a change in meaning
• Ex) wide (adjective) -gt widely (adverb)
• Compounding
• The merging of two or more words into a new word
• Ex) noun-noun compounds (tea kettle, disk drive)

7
Nouns and pronouns (1/2)

• Nouns
• refer to entities in the world like people,
  animals, and things
• (3.3)
• Table 3.1 Common inflections of nouns
• The plural form
• English has only one inflection of the noun
• Suffix s
• regular vs irregular
• (3.4)

8
Nouns and pronouns (2/2)

• Pronouns
• A class of words that act like variables in that
  they refer to a person or thing that is somehow
  salient in the context
• (3.5)
• Table 3.2 Pronoun forms in English

9
Determiners and adjectives

• Determiners
• describe the particular reference of a noun
• Articles
• the, a (or an)
• Demonstratives
• this, that
• Adjectives
• describe properties of nouns
• (3.7) attributive or adnominal adjectives
• (3.8) predicative adjectives
• Comparative (-er) vs superlative (-est)
• Periphrastic form using more, most

10
Verbs (1/2)

• Verbs
• Used to describe actions, activities and states
• Regular verb has the following morphological
  forms
• The root or base form walk
• The third singular present tense walks
• The gerund and present participle walking
• The past tense form and past/passive participle
  walked

• Base form
• Present tense
• Third singular person
• Infinitive with to
• Bare infinitive
• -ing form
• Progressive
• Gerund
• -ed form
• Past
• Present perfect
• Past perfect

11
Verbs (2/2)

• Irregular verb
• Have different forms for past tense and past
  participle
• Ex) drive-drove-driven, take-took-taken
• Table 3.3 Features commonly marked on verbs
• Synthetic forms -s, -ing, -ed
• Analytic forms with auxiliary
• Auxiliary
• Words that accompany verbs
• Used to express aspect, mood, and some tense
  information
• have, be, will, may, can, should

12
Other parts of speech (2/2)

• Adverbs
• Modify a verb
• Specify place, time, manner or degree
• Some can modify adjectives and other adverbs
• Prepositions
• Express spatial relationship
• (3.24)
• Particles
• A subclass of prepositions that can enter into
  strong bonds with verbs in the formation of
  so-called phrasal verbs.
• (3.25), (3.26)
• Need to be able to distinguish particles and
  prepositions
• (3.27)

13
Other parts of speech (2/2)

• Conjunctions
• Coordinating conjunctions
• conjoin or coordinate two words or phrases of
  the same category
• and, or, but
• Subordinating conjunctions
• attaches a secondary sentence to a primary
  sentence
• (3.28)

14
Phrase Structure

• Syntax
• The study of the regularities and constraints of
  word order and phrase structure
• Constituent
• Phrases behave as constituents
• Can be detected by their being able to occur in
  various positions, and showing uniform syntactic
  possibilities for expansion
• (3.29), (3.30)
• Some of the major phrase types
• Noun phrases
• Prepositional phrases
• Verb phrases
• Adjective phrases

15
Phrase structure grammars

• Syntactic analysis
• How to determine the meaning of the sentence from
  the meaning of the words
• word order allows us to infer who did what to
  whom
• (3.34), (3.35)
• Basic word order Subject-Verb-Object
• (3.36)
• In questions or commands or requests, this order
  is modofied
• (3.37), (3.38)

16
Phrase structure grammars

• Rewrite rule
• Capture the regularities of word order
• The form category -gt category
• The symbol (left) can be rewritten as the
  sequence of symbols (right)
• To produce a sentence, we start with the start
  symbol S
• (3.39)
• Context-free grammar
• The possibilities for rewriting depend on the
  category, and not on any surrounding context

17
Phrase structure grammars

• To represent phrase structure
• Derivation
• (3.40), (3.41)
• Tree
• (3.42), (3.43)
• (labeled) bracketing
• (3.44)
• Recursivity
• makes it possible for a single nonterminal symbol
  to be extended to a large number of words
• Ex) NP -gt NP PP PP -gt IN NP
• Figure 3.1

18
Phrase structure grammars

• Non-local dependency
• Two words that are syntactically linked can
  become separated by intervening words
• Because two words can be syntactically dependent
  even though they occur far apart in a sentence
• Subject-verb agreement (3.45)
• Long-distance dependency (3.46)
• A challenge for some Statistical NLP approaches
  like n-grams
• Empty node
• When nonterminal may be rewritten as nothing
• Ex) Eat the cake! NP -gt e

19
Dependency

• Subject Children (the agents of the action of
  eating)
• Object sweet candy (the patient of the action)
• Arguments of verb eat children, sweet candy
• Indirect object him
• Direct object the book

Children eat sweet candy
She gave him the book
She gave the book to him
20
Dependency

• Active voice
• Subject Children (the agents of the action of
  eating)
• Object sweet candy (the patient of the action)
• Passive voice
• Subject Candy (the patient of the action of
  eating)
• By-phrase children (the agents of the action)

Children eat sweet candy
Candy is eaten by children
21
Dependency

• The dependents of verbs
• Arguments
• express entities that are centrally involved in
  the activity of the verb
• Obligatory
• subject, complement
• Adjuncts
• phrases that have a less tight link to the verb
• Optional
• The time, place, or manner of the action or state
• (3.57)

22
X theory

• ?? ?????? ??
• ????? ????? ????? ????? ???? ????? ?? ??
• ???? ??(? NP-gtV VP)? ???? ??? ?? ??
• ? ??? ?? ?? ?? ??? ??
• X theory
• ???? ???? ?? ??
• A word will be the head of a phrase
• A head forms a small constituent with its
  complements
• This constituent can be modified by adjuncts to
  form a bigger constituent
• This constituent can combine with a specifier

23
X theory
Head(???) Complements(???) sisters of X dominated
by X Adjuncts(???) sisters of X dominated by
X Specifier(???) sister of X dominated by X
24
Phrase structure ambiguity

• Parsing
• The process of reconstructing the derivation(s)
  or phrase structure tree(s) that give rise to a
  particular sequence of words
• Parse phrase structure tree
• Phrase structure ambiguity (syntactic ambiguity)
• There are many different parse trees that could
  all have given rise to a particular sequence of
  words
• (1.10) Our company is training workers
• Attachment ambiguity
• Occur with phrases that could have been generated
  by two different nodes
• (3.65) The children ate the cake with a spoon

25
Phrase structure ambiguity

• Garden pathing
• The phenomenon of first being tricked into
  adopting a spurious parse and then having to
  backtrack to try to construct the right parse
• (3.66) The horse raced past the barn fell
• Rare in spoken language
• A real problem when reading complex sentences of
  written English
• There is no parse
• Sentence that is not covered by the grammar
• The sentence is ungrammatical
• Ungrammaticality (3.67) vs semantic abnormality
  (3.68)

26
Semantics and Pragmatics (1/3)

• Semantics
• The study of the meaning of words, constructions,
  and utterances
• Two parts
• The study of the meaning of individual words
  (lexical semantics)
• The study of how meanings of individual words are
  combined into the meaning of sentences
• Approach to lexical semantics
• To study how word meanings are related to each
  other
• The principle of compositionality
• The meaning of the whole can be strictly
  predicted from the meaning of the parts
• Natural language often does not obey the
  principle
• (3.69)

27
Semantics and Pragmatics (2/3)

• Terminology
• Hypernym(hyperonym) vs Hyponym
• Antonym
• Meronym vs Holonym
• Synonym
• Homonym
• Polyseme
• Homophony
• Idiom
• A phrase that relationship between the meaning of
  the words and the meaning of the phrase is
  completely opaque
• kick the bucket dying
• The problem of scope
• (3.70)

28
Semantics and Pragmatics (3/3)

• Discourse
• The study of the covert relationships between
  sentences in a text
• the resolution of anaphoric relations
• Between noun phrases that refer to the same
  person or thing
• A central problem in discourse analysis
• (3.71)
• is important for information extraction
• (3.72)
• Pragmatics
• The study of how world knowledge and language
  conventions interact with literal meaning