Arabic Dialect Syntax and Tree Adjoining Grammar

1
Arabic Dialect Syntax andTree Adjoining Grammar

• Owen Rambow
• Columbia University
• rambow_at_cs.columbia.edu

2
Overview

• Morphology and Syntax
• Phrase Structure for MSA
• Dialect Syntax
• Parsing Dialect Syntax
• Tree Adjoining Grammar

3
Morphology and Syntax

• Rich morphology crosses into syntax
• Pro-drop / Subject conjugation
• Verb subcategorization and subject/object clitics
• Verbtransitivesubjectobject
• Verbintransitivesubject but not
  Verbintransitivesubjectobject
• Verbtransitivepassivesubject but not
  Verbtransitivepassivesubjectobject
• Verbintransitivepassive but not
  Verbintransitivepassivesubject

4
Morphology and Syntax

• Morphological interactions with syntax
• Agreement
• Full e.g. Noun-Adjective on number, gender, and
  definiteness
• Partial e.g. Verb-Subject on gender (in VSO
  order)
• Definiteness
• Noun compound formation, copular sentences, etc.
• NounsDefiniteArticle, Proper Nouns, Pronouns,
  etc.

5
Morphology and Syntax

• Morphological interactions with syntax
  (continued)
• Case
• MSA is case marking nominative, accusative,
  genitive
• Almost-free word order
• Case is often marked with optionally written
  short vowels
• This effectively limits the word-order freedom in
  published text
• Agglutination
• Attached prepositions create words that cross
  phrase boundaries
• ????????? liAlmaktabat
• for the-libraries PP li NP Almaktabat
• Some morphological analysis (minimally
  segmentation) is necessary even for statistical
  approaches to parsing

6
Sentence Structure

• Traditional Arabic Grammar Two types of Arabic
  Sentences
• Verbal sentences
• Verb Subject Object (VSO)
• ??? ??????? ???????Wrote the-boys the-poemsThe
  boys wrote the poems
• Copular sentences
• Topic Complement
• ??????? ?????the-boys poetsThe boys are poets

7
Verbal Sentences

• Verb agreement in VSO with gender only
• ??? ?????\??????? wrote3MascSing the-boy/the-boys
• ???? ?????\?????? wrote3FemSing
  the-girl/the-girls
• Pronominal subjects are cliticized
• ????? wrote-youMascSing
• ????? wrote-youMascPlur
• ????? wrote-theyMascPlur

8
VSO vs SVO vs OVS

• ???? ?????????????wrote.fem the-girls
  the-poemsThe girls wrote the poems
• ???????????wrote-they.fem the-poemsThey.fem
  wrote the poems
• ?????? ???? ???????the-girls wrote-they.fem
  the-poemsThe girls wrote the poems
• ??????? ?????? ?????? the-poems wrote.fem-them
  the girlsThe poems, the girls wrote them

9
VSO, VOS, SVO, OVSDescriptive Generalization

• VSO or VOS Agreement with subject in gender only
• Subject pronoun is a clitic on verb and replaces
  agreement
• SVO order has preposed subject followed by verb
  with subject clitic
• Object pronoun is a clitic on verb (does not
  replace subject agreement)
• OVS order has preposed object followed by verb
  with object clitic

10
VSO Phrase Structure

• English vanilla case

S
NP
VP
the boys
V
NP
the poems
wrote
11
VSO Phrase Structure

• VSO Option 1

Penn Arabic Treebank
S
VP
NP
V
NP
the poems
wrote
the boys
12
VSO Phrase Structure

• VSO Option 2

S
NP
VP
the boys
V
NP
the poems
wrote
13
VSO Phrase Structure

• VSO Option 2

S?
S
V
wrote
NP
VP
the boys
V
NP
?
the poems
14
VSO Phrase Structure

• VSO Option 2

S?
S
V
Head Movement
wrote
NP
VP
the boys
V
NP
?
the poems
15
SVO Phrase Structure

• Option 1 English phrase structure

S
Problem Arabic does not look like English
(subject clitic on verb)
NP
VP
the boys
V
NP
the poems
wrote
16
SVO Phrase Structure

• Option 2

S
VP
NP
V
NP
the poems
wrote
the boys
17
SVO Phrase Structure

• Option 2

Penn Arabic Treebank
S
NP
VP
the boys
V
NP
NP
the poems
wrote
?
18
SVO Phrase Structure
S?

• Option 3

NP
S?
the boys
S
V
wrote
NP
VP
?
V
NP
the poems
?
19
Copular sentences

• Topic Complement
• Definite Topic, Indefinite Complement
• ????? ????the-boy poetThe boy is a poet
• Auxiliary Topic Complement
• Auxiliaries (kana and her sisters)
• Tense, Negation, Transformation, Persistence
• ??? ????? ????? was the-boy poet The boy was a
  poet
• ??? ????? ????? is-not the-boy poet The boy is
  not a poet

20
Copular Sentences

• Types of complements
• Noun/Adjective/Adverb
• ????? ??? the-boy smart The boy is smart
• Prepositional Phrase
• ????? ?? ??????? the-boy in the-library The boy
  is in the library

21
SVO, OVS as Copular Sentence

• Verb-Sentence
• ??????? ????? ???????
• the-boys wrote-they poems The boys wrote
  the poems
• ??????? ?????? ??????
• the-poems wrote.fem-them the girls The
  poems, the girls wrote them
• Copular-Sentence
• ????? ????? ???? the-boy book-his big The
  boy, his book is big

22
Common Structural Ambiguities

• Third masculine/feminine singular are
  structurally ambiguous
• Verb3MascSingular NounMasc
• Verb subjecthe objectNoun
• Verb subjectNoun
• Passive and active forms are often similar in
  standard orthography
• ??? /kataba/ he wrote
• ???? /kutiba/ it was written

23
Overview

• Morphology and Syntax
• Phrase Structure for MSA
• Dialect Syntax
• Parsing Dialect Syntax
• Tree Adjoining Grammar

24
Dialect Syntax

• MSA
• Verb Subject Object ??? ??????? ???????
• wrote.masc the-boys the-poems
• Subject Verb Object (Full agreement)
• ??????? ????? ???????
• the-boys wrote-they.masc the-poems
• LEV, EGY
• Subject Verb Object
• ??????? ???? ???????
• The-boys wrote.masc.pll the-poems
• Less frequent Verb Subject Object
• ???? ??????? ???????
• wrote.masc.pl the-boys the-poems
• Full agreement (or clitic?) in both orders

25
Dialect Syntax Noun Phrases

• Possessives
• Idafa construction
• Noun1 of Noun2 encoded structurally
• ??? ??????
• king Jordan
• the king of Jordan / Jordans king
• Dialects have an additional common construct
• Noun1 ltparticlegt Noun2
• LEV ????? ??? ?????? the-king belonging-to
  Jordan
• ltparticlegt differs widely among dialects
• Pre/post-modifying demonstrative article
• MSA ??? ????? this the-man this man
• EGY ?????? ?? the-man this this man

26
Code Switching

• MSA and Dialect mixing in speech
• phonology, morphology and syntax

MSA LEV

• ?? ??? ?? ????? ???? ????? ???? ?? ???????? ?????
  ????? ?????? ???? ?? ???? ?????? ???????? ??????
  ??????? ???????? ????? ??? ????? ????? ??? ?????
  ??? ????? ??? ???? ?? ???? ????????? ?????? ????
  ???? ?? ?????? ????? ??????????? ??? ???? ??
  ?????? ????????? ?????? ??? ???? ?? ????? ??
  ?????? ????? ?? ????? ???? ??? ???????? ?? ???
  ???? ???? ??? ????? ??????? ????? ???? ??? ????
  ?? ??????? ????? ??? ?? ?????? ?? ????? ????
  ????? ?????? ?? ????? ?? ??? ?????? ??? ????
  ????? ???????? ?????? ?? ????? ??? ???????
  ?????? ??????? ???? ???? ???? ??????? ???????
  ???? ??? ????? ?? ??? ????? ?? ???? ???? ???? ???
  ??? ??????? ????? ???????? ?? ????? ????????? ???
  ????? ????? ????? ??? ???? ?????? ???? ????? ??
  ??? ????? ???? ?? ????? ?? ???? ??????? ?? ????
  ???? ?????? ????????? ???? ??? ??? ?? ????? ??
  ??? ????? ?????? ???? ?????? ????????? ????
  ??????? ???? ????? ????????? ???? ????? ?? ?? ???
  ??? ?? ?? ???? ????? ???? ??????? ??????? ?? ???
  ?? ?????? ????? ?? ??? ?? ????? ?? ??? ??????
  ???????? ?? ????? ????? ????? ??? ????? ?? ????
  ?????? ???? ?????? ????? ??? ???? ???? ????? ???
  ????? ????? ???? ?? ????? ???? ???? ????? ???
  ?????? ???? ??????? ???? ??? ???? ???? ??????
  ????? ????????? ???????? ??? ?????? ???? ????
  ??????? ???? ??????? ??? ?????? ???? ??? ????? ??
  ??? ???????? ??? ??????? ?????????? ??? ?????
  ????? ??? ??????? ????? ?? ?? ???? ???? ???
  ??????? ?? ?????? ?? ?? ??????? ??? ????? ??????
  ???????? ??? ?????? ???????? ??? ?? ????? ?????
  ??????? ?????? ????? ?? ??? ???? ?? ????
  ??????????? ??? ?????? ???? ?????? ?? ??? ???????.

Aljazeera Transcript http//www.aljazeera.net/prog
rams/op_direction/articles/2004/7/7-23-1.htm
27
Parsing Arabic DialectsProblem
Dialect
MSA
Treebank
??????? ???? ???????
?
????
Parser
???????
???????
28
Parsing Arabic Dialects

• Many different dialects
• Dialects are spoken, few written resources
• Code switching
• Conclusion Cant assume we will get treebanks
  for each dialect
• What to do?

29
Parsing Arabic Dialects

• Idea use resources for MSA, apply them to
  dialects
• We will be investigating three approaches

30
Parsing Arabic DialectsProposed Solution 1
Dialect
MSA
Treebank
??????? ???? ???????
??? ??????? ???????
Parser
31
Parsing Arabic DialectsProposed Solution 2
Dialect
MSA
Treebank
??????? ???? ???????
Parser
32
Parsing Arabic DialectsProposed Solution 3
Dialect
MSA
Treebank
??????? ???? ???????
Parser
33
Overview

• Morphology and Syntax
• Phrase Structure for MSA
• Dialect Syntax
• Parsing Dialect Syntax
• Tree Adjoining Grammar

34
(No Transcript)