prague dependency treebank - Úfalufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf ·...
TRANSCRIPT
![Page 1: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/1.jpg)
Prague Dependency Treebank: Introduction – trees, dependency
Markéta Lopatková, Jiří Mírovský
Institute of Formal and Applied Linguistics, MFF UK [email protected]
![Page 2: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/2.jpg)
Lectures: Markéta Lopatková Fri, S6, 14:00-15:30
Practical sessions: Jiří Mírovský Fri, SU1, 12:20-13:50
http://ufal.mff.cuni.cz/course/npfl075
PDT – Intro Lopatková
NPFL075 Prague Dependency Treebank
Requirements: • Homework (40%) • Activity (10%) • Final test (50%)
Assessment: • excellent (= 1) ≥ 90% • very good (= 2) ≥ 70% • good (= 3) ≥ 50%
![Page 3: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/3.jpg)
Prague Dependency Treebank
PDT – Intro Lopatková
Collection of: • linguistically annotated data (Czech) • tools and data format(s) • documentation
Another point of view: • annotation scheme • framework for annotation of different languages • underlying linguistic theory (Functional Generative Description)
![Page 4: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/4.jpg)
Collection of: • linguistically annotated data (Czech) • tools and data format(s) • documentation
Another point of view: • annotation scheme • framework for annotation of different languages • underlying linguistic theory (Functional Generative Description)
What about other/similar approaches: • HamleDT • Universal Dependencies
Prague Dependency Treebank
PDT – Intro Lopatková
![Page 5: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/5.jpg)
Outline of the lecture
trees (graph theory and data format) phrase structure trees and dependency trees dependency and non-dependency relations non-projectivity
PDT – Intro Lopatková
![Page 6: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/6.jpg)
PDT – Intro Lopatková
How to capture sentence structure?
![Page 7: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/7.jpg)
Graph theory: tree tree (graph theory): definiftion:
finite graph ⟨N, E⟩, N ~ nodes/vertices, E ~ edges {n1,n2} connected no cycles, no loops no more than 1 edge between any two different nodes ⇔ (undirected) graph
any two nodes are connected by exactly one simple path
PDT – Intro Lopatková
![Page 8: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/8.jpg)
Graph theory: tree tree (graph theory): definiftion:
finite graph ⟨N, E⟩, N ~ nodes/vertices, E ~ edges {n1,n2} connected no cycles, no loops no more than 1 edge between any two different nodes ⇔ (undirected) graph
any two nodes are connected by exactly one simple path rooted tree
rooted ⇒ orientation (i.e., edges ordered pairs [n1,n2])
PDT – Intro Lopatková
![Page 9: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/9.jpg)
Graph theory: tree tree (graph theory): definiftion:
finite graph ⟨N, E⟩, N ~ nodes/vertices, E ~ edges {n1,n2} connected no cycles, no loops no more than 1 edge between any two different nodes ⇔ (undirected) graph
any two nodes are connected by exactly one simple path rooted tree
rooted ⇒ orientation (i.e., edges ordered pairs [n1,n2])
directed tree … directed graph which would be tree - if the directions on the edges were ignored, or - all edges are directed towards a particular node ~ the root
PDT – Intro Lopatková
![Page 10: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/10.jpg)
Data structure: tree tree as a data structure:
rooted tree (as in graph theory) all edges are directed from a particular node ~ the root
PDT – Intro Lopatková
![Page 11: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/11.jpg)
Data structure: tree tree as a data structure:
rooted tree (as in graph theory) all edges are directed from a particular node ~ the root + (linear) ordering of nodes: the children of each node have a specific order
PDT – Intro Lopatková
![Page 12: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/12.jpg)
tree as a data structure: "tree-ordering" D … partial ordering on nodes
u ≤ v ⇔def the unique path from the root to v passes through u (weak ordering ~ reflexive, antisymmetric, transitive)
"linear ordering" … (partial) ordering on nodes (strong ordering ~ antireflexive, asymmetric, transitive)
Data structure: tree (properties)
PDT – Intro Lopatková
![Page 13: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/13.jpg)
Tree-based structures in CL two types of tree-based structures in CL: phrase structure tree / constituent structure tree dependency tree
PDT – Intro Lopatková
![Page 14: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/14.jpg)
My brother often sleeps in his study.
Phrase structure tree
study
S
NP VP
Det NP
my
brother
VP
V PP
Prep
sleeps in
Adv
often
NP
N
Det NP
his N Noam Chomsky (1957) Syntactic Structures. The Hague: Mouton
PDT – Intro Lopatková
![Page 15: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/15.jpg)
Phrase structure tree (definition)
T = ⟨ N, D, Q, P, L ⟩ ⟨N, D⟩ … rooted tree Q ... lexical and grammatical categories L … labeling function N → Q D … oriented edges (branches) ~ relation on lex. and gram. categories dominance relation + P ... relation on N ~ (partial strong linear ordering) relation of precedence
PDT – Intro Lopatková
![Page 16: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/16.jpg)
Phrase structure tree (definition)
Relating dominance and precedence relations: exclusivity condition for D and P relations ‘nontangling’ condition
PDT – Intro Lopatková
+
T = ⟨ N, D, Q, P, L ⟩ ⟨N, D⟩ … rooted tree, directed Q ... lexical and grammatical categories L … labeling function N → Q D … oriented edges (branches) ~ relation on lex. and gram. categories dominance relation + P ... relation on N ~ (partial strong linear ordering) relation of precedence
![Page 17: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/17.jpg)
exclusivity condition for D and P relations ∀ x,y∈ N holds: ( [x,y] ∈ P ∨ [y,x] ∈ P ) ⇔ ( [x,y] ∉ D & [y,x] ∉ D)
Phrase structure tree (relation P)
S
NP VP
Det NP
my
brother
VP
V PP
Prep
sleeps in
Adv
often
NP
N
Det NP
his N
study
PDT – Intro Lopatková
![Page 18: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/18.jpg)
exclusivity condition for D and P relations ∀ x,y∈ N holds: ( [x,y] ∈ P ∨ [y,x] ∈ P ) ⇔ ( [x,y] ∉ D & [y,x] ∉ D) ‘nontangling’ condition ∀ w,x,y,z∈ N holds: ( [w,x] ∈ P & [w,y] ∈ D & [x,z] ∈ D ) ⇒ ( [y,z] ∈ P )
Phrase structure tree (relation P)
w x
y z
w x
y z
w x
y z
w x
y z
PDT – Intro Lopatková
![Page 19: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/19.jpg)
exclusivity condition for D and P relations ∀ x,y∈ N holds: ( [x,y] ∈ P ∨ [y,x] ∈ P ) ⇔ ( [x,y] ∉ D & [y,x] ∉ D) ‘nontangling’ condition ∀ w,x,y,z∈ N holds: ( [w,x] ∈ P & [w,y] ∈ D & [x,z] ∈ D ) ⇒ ( [y,z] ∈ P )
Phrase structure tree (relation P)
T = ⟨ N,D,Q, P,L ⟩ phrase structure tree ∀ x,y ∈ N siblings ⇒ [x,y ] ∈ P the set of its leaves is totally ordered by P
PDT – Intro Lopatková
![Page 20: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/20.jpg)
Phrase structure tree
Pros • derivation history / ‘closeness’ of a
complementation • coordination, apposition • CFG-like • derivation of a grammar
PDT – Intro Lopatková
study
S
NP VP
Det NP
my
brother
VP
V PP
Prep
sleeps in
Adv
often
NP
N
Det NP
his N
![Page 21: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/21.jpg)
Phrase structure tree
study
VP
VP
V PP
Prep
sleeps in
Adv
often
NP
Det NP
his N
VP
V
PP
Prep
sleeps
in
VP
often
NP
Det NP
his
study
N
Adv
… often sleeps in his study … often sleeps in his study
derivation history / ‘closeness’:
PDT – Intro
![Page 22: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/22.jpg)
Phrase structure tree
Pros • derivation history / ‘closeness’ of a
complementation • coordination, apposition • CFG-like • derivation of a grammar
Contras • complexity (number of non-terminal symbols) • complement (‘two dependencies’) přiběhl bos [(he) arrived barefooted] • free word order discontinuous ‘phrases’ non-projectivity
PDT – Intro Lopatková
![Page 23: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/23.jpg)
Phrase structure tree
PDT – Intro Lopatková
S
NP
Mary
VP
VP
AuxV
what
will
V
N NP
eat
N
S
NP
Mary
VP
VP
AuxV
bread will
V
N NP
eat
N
discontinuous ‘phrases’: solution for English Mary will eat bread. What will Mary eat?
![Page 24: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/24.jpg)
discontinuous ‘phrases’: solution for English Mary will eat bread. What will Mary eat?
Phrase structure tree
S
NP
Mary
VP
VP
AuxV
bread will
V
N NP
eat
N
S
NP
Mary
VP
VP
AuxV
what
will
V
N NP
eat
N
PDT – Intro Lopatková
![Page 25: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/25.jpg)
discontinuous ‘phrases’: solution for English Mary will eat bread. What will Mary eat?
Phrase structure tree
S
NP
Mary
VP
VP
AuxV
bread will
V
N NP
eat
N
S
NP
Mary
VP
VP
AuxV
tracej
tracei
V
N NP
eat
N
T'
AuxV
will
S'
NP
what
PDT – Intro Lopatková
![Page 26: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/26.jpg)
Po babiččině příjezdu půjdou rodiče do divadla. [After grandma's arrival the parents will go to the theatre.]
S
VP NP
PrepP VP
Prep
půjdou
N
rodiče NP
Atr N
příjezdu babičině
V po
VP PrepP
Prep NP
do N
divadla
Phrase structure tree discontinuous ‘phrases’:
PDT – Intro Lopatková
![Page 27: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/27.jpg)
Corpora with phrase structure trees
• Penn Treebank (1995) Mitchel Marcus (1993) Computational Linguistics, vol. 19 http://www.cis.upenn.edu/~treebank/ Penn Arabic Treebank, Penn Chinese Treebank • International English Treebank (ICE) http://ice-corpora.net/ice/index.htm • Paris 7 http://www.llf.cnrs.fr/Gens/Abeille/French-Treebank-fr.php • Szeged Treebank 2.0 http://www.inf.u-szeged.hu/projectdirs/hlt/en/Szeged%20Treebank%202.0_en.html • many many others
PDT – Intro Lopatková
![Page 28: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/28.jpg)
Dependency tree
PDT – Intro Lopatková
![Page 29: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/29.jpg)
My brother often sleeps in his study.
sleeps.Pred
brother.Sb
study.Adv my.Atr
often.Adv
his.Atr
in.AuxP
Dependency tree
Lucien Tesnière (1959) Éléments de syntaxe structurale. Editions Klincksieck. Igor Mel’čuk (1988) Dependency Syntax: Theory and Practice. State University of New York Press.
My brother often sleeps in his study.
PDT – Intro Lopatková
![Page 30: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/30.jpg)
Dependency tree (definition) T = ⟨ N, D, Q, WO, L ⟩ ⟨N, D⟩ … rooted tree, directed Q ... lexical and grammatical categories L … labeling function N → Q D … oriented edges ~ relation on lex. and gram. categories ‘dependency’ relation WO ...relation on N ~ (strong total ordering on N) … word order
PDT – Intro Lopatková
sleeps.Pred
brother.Sb
study.Adv my.Atr
often.Adv
his.Atr
in.AuxP
![Page 31: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/31.jpg)
Dependency tree
Pros • economical, clear (complex labels, ‘word’~ node) • free word order • head of a phrase
Contras • no derivation history / 'closeness' • coordination, apposition • complement
PDT – Intro Lopatková
sleeps.Pred
brother.Sb
study.Adv my.Atr
often.Adv
his.Atr
in.AuxP
![Page 32: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/32.jpg)
Dependency tree
PDT – Intro Lopatková
eat.Pred
bread.Obj Mary.Sb will.AuxV
discontinuous ‘phrases’: no problem Mary will eat bread. What will Mary eat?
eat.Pred
Mary.Sb What.Obj will.AuxV
![Page 33: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/33.jpg)
Dependency tree Po babiččině příjezdu půjdou rodiče do divadla. [After grandma's arrival the parents will go to the theatre.]
půjdou.Pred
příjezdu.Adv
rodiče.Sb
babiččině.Atr
do.AuxP po.AuxP
divadla.Adv
PDT – Intro Lopatková
![Page 34: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/34.jpg)
Corpora with dependency trees • PropBank (1995) http://propbank.github.io/ • family of Prague dependency treebanks: Czech, Arabic, English http://ufal.mff.cuni.cz/pdt.html • HamleDT project (from 2012) http://ufal.mff.cuni.cz/hamledt • Universal Dependencies http://universaldependencies.org/
• Danish Dep. Treebank http://mbkromann.github.io/copenhagen-dependency-treebank/ • Finnish: Turku Dependency Treebank http://bionlp.utu.fi/fintreebank.html • Negra corpus http://www.coli.uni-saarland.de/projects/sfb378/negra-corpus/negra-corpus.html • TIGERCorpus http://www.ims.uni-stuttgart.de/forschung/ressourcen/korpora/tiger.html/ • SynTagRus Dependency Treebank for Russian
PDT – Intro Lopatková
![Page 35: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/35.jpg)
Dependency and non-dependency relations
PDT – Intro Lopatková
![Page 36: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/36.jpg)
edges ~ dependency relations (prototypically) • dependency relation: binary relation • governing/modified unit (head) – dependent/modifying unit (modifier) • long discussion, number of linguistic criteria e.g., each complete subtree must be a “constituent“, i.e., it must allow for several
constructions like topicalization, proform substitution, ….;
Dependency and non-dependency relations
PDT – Intro Lopatková
will.Pred
bread.Obj Mary.Sb eat.???
Topicalization: … and eat Mary certainly will.
Proform substitution: Mary will do so. (do=eat)
Answer fragment: What will Mary do? Eat.
VP-ellipsis: Peter will eat and Mary will, too. lexical verb should be a dependent
Mary will eat bread..
![Page 37: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/37.jpg)
edges ~ dependency relations (prototypically) • dependency relation: binary relation • governing/modified unit (head) – dependent/modifying unit (modifier)
• PDT criterion: possible reduction … dependent member of the pair may be deleted while the distributional properties are preserved (→ correctness is preserved)
Dependency and non-dependency relations
PDT – Intro Lopatková
![Page 38: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/38.jpg)
edges ~ dependency relations (prototypically) • dependency relation: binary relation • governing/modified unit (head) – dependent/modifying unit (modifier)
• PDT criterion: possible reduction … dependent member of the pair may be deleted while the distributional properties are preserved (→ correctness is preserved)
• endocentric constructions … OK malý stůl stůl small table table přišel včas přišel he came in time he came (přišel) velmi brzo (přišel) brzo (he came) very soon (he came) soon
Dependency and non-dependency relations
PDT – Intro Lopatková
![Page 39: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/39.jpg)
edges ~ dependency relations (prototypically) • dependency relation: binary relation • governing/modified unit (head) – dependent/modifying unit (modifier)
• PDT criterion: possible reduction … dependent member of the pair may be deleted while the distributional properties are preserved (→ correctness is preserved)
• endocentric constructions … OK
• exocentric constructions … principle of analogy on word classes
Prší. [(It) rains.] … ∃ subjectless verbs ⇒ Král zemřel. [The king died.] … a verb rather than a noun is the head The girl painted a bag. → The girl painted. ... ∃ objectless verbs ⇒ The girl carried a bag … an object is considered as depending on a verb
Dependency and non-dependency relations
PDT – Intro Lopatková
![Page 40: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/40.jpg)
edges ~ dependency relations (prototypically) • dependency relation: binary relation • governing/modified unit (head) – dependent/modifying unit (modifier)
• PDT criterion: possible reduction … dependent member of the pair may be deleted while the distributional properties are preserved (→ correctness is preserved)
• endocentric constructions … OK
• exocentric constructions … principle of analogy on word classes
PLUS technical considerations e.g.: prepositions are below nouns; auxiliary verbs are (typically) below content verbs
Dependency and non-dependency relations
PDT – Intro Lopatková
![Page 41: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/41.jpg)
Dependency and non-dependency relations BUT also other relations: coordination … "multiplication" of a single syntactic position different referents coordination of sentence members / sentences My sister Mary and John came late. Mary came in time but John was late. I can't leave since it hasn't stopped raining yet. Nemohu odejít, neboť ještě nepřestalo pršet. coordination may be embedded
nice and romantic towers and castles krásné a romantické hrady a zámky
PDT – Intro Lopatková
![Page 42: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/42.jpg)
BUT also other relations: coordination … "multiplication" of a single syntactic position different referents coordination of sentence members / sentences My sister Mary and John came late. Mary came in time but John was late. I can't leave since it hasn't stopped raining yet. Nemohu odejít, neboť ještě nepřestalo pršet. coordination may be embedded nice and romantic towers and castles krásné a romantické hrady a zámky
apposition … "multiplication" of a single syntactic position identical referent Charles IV, Holy Roman Emperor The Hobbit, or There and Back Again
Dependency and non-dependency relations
PDT – Intro Lopatková
![Page 43: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/43.jpg)
BUT also other relations: coordination … "multiplication" of a single syntactic position different referents coordination of sentence members / sentences coordination may be embedded
apposition … "multiplication" of a single syntactic position identical referent
necessary to enrich the data structure
Dependency and non-dependency relations
PDT – Intro Lopatková
![Page 44: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/44.jpg)
PDT – Intro
Coordination/apposition in dependency trees PDT 2.0: 'connecting' constructions ~ coordination, apposition (, OPER)
specific types of nodes and edges: connecting node (= node for coordinating / appositing conjunction)
men Sb_Co
came Pred
Thin Atr
soldiers Sb_Co
and Coord
young Atr
![Page 45: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/45.jpg)
PDT – Intro
Coordination/apposition in dependency trees PDT 2.0: 'connecting' constructions ~ coordination, apposition (, OPER)
specific types of nodes and edges: connecting node (= node for coordinating / appositing conjunction) effective parent (= node for governing node, i.e. node modified by the whole construction, 'linguistic parent')
men Sb_Co
came Pred
Thin Atr
soldiers Sb_Co
and Coord
young Atr
![Page 46: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/46.jpg)
PDT – Intro
Coordination/apposition in dependency trees PDT 2.0: 'connecting' constructions ~ coordination, apposition (, OPER)
specific types of nodes and edges: connecting node (= node for coordinating / appositing conjunction) effective parent (= node for governing node, i.e. node modified by the whole construction, 'linguistic parent') members of a connecting construction (= nodes that are coordinated / are in apposition)
is_member
men Sb_Co
came Pred
Thin Atr
soldiers Sb_Co
and Coord
young Atr
![Page 47: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/47.jpg)
PDT – Intro
Coordination/apposition in dependency trees PDT 2.0: 'connecting' constructions ~ coordination, apposition (, OPER)
specific types of nodes and edges: connecting node (= node for coordinating / appositing conjunction) effective parent (= node for governing node, i.e. node modified by the whole construction, 'linguistic parent') members of a connecting construction (= nodes that are coordinated / are in apposition)
is_member effective child(ren) … modification(s) of the individual member of the connecting construction + common/shared modifier(s) ‘pass-through’ nodes men
Sb_Co
came Pred
Thin Atr
soldiers Sb_Co
and Coord
young Atr
![Page 48: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/48.jpg)
The center will gather and distribute the information on tenders and state commissions in this country as well as in abroad.
![Page 49: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/49.jpg)
Coordination/apposition in dependency trees PDT 2.0: embedded connecting constructions recursivity TrEd (Tree Editor, Pajas): functions GetEChildren, GetEParents
PDT – Intro Lopatková
![Page 50: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/50.jpg)
Universal Dependencies: version 1 (2014): the first conjunct ~ the head of all following conjuncts ~ the head of any intervening coordinating conjunctions and punctuation
Coordination/apposition in dependency trees
PDT – Intro Lopatková
![Page 51: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/51.jpg)
Universal Dependencies: version 1 (2014): the first conjunct ~ the head of all following conjuncts ~ the head of any intervening coordinating conjunctions and punctuation version 2 (2016): • the first conjunct ~ the head of all following conjuncts • attach coordinating conjunctions and punctuation to the immediately succeding conjunct (instead of the first)
Coordination/apposition in dependency trees
PDT – Intro Lopatková
![Page 52: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/52.jpg)
Mel'čuk (1988): • ‘grouping’ (G) … treating the first conjunct as the head • problem: shared modification vs. modification of a single member
Coordination/apposition in dependency trees
Hubení ( ( mladí muži ) , vojáci a starci ) [Thin young men, soldiers and old-men]
PDT – Intro Lopatková
![Page 53: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/53.jpg)
Coordination/apposition in dependency trees
PDT – Intro Lopatková
Petkevič (1995) … formal representation of FGD two types of brackets for tree linearization: • ⟨ ⟩ for dependencies • [ ] for coordination
![Page 54: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/54.jpg)
PDT – Intro Lopatková
References • Hajičová, E., Havelka, J., Sgall, P., Veselá, K., Zeman, D. (2004) Issues of
Projectivity in the Prague Dependency Treebank. PBML, vol. 81 • Holan, T., Kuboň, V., Oliva, K., Plátek, M. (2000) On Complexity of Word
Order. Les grammaires de dépendance – Traitement automatique des langues, vol. 41, no. 1, 273-300
• Kuhlmann, M., Nivre, J. (2006) Mildly Non-Projective Dependency Structures. In COLING/ACL Main Conference Poster Sessions, 507–514.
• Mel’čuk, I. (1988) Dependency Syntax: Theory and Practice. State University of New York Press, Albany
• Partee, B. H.; ter Meulen, A.; Wall, R. E. (1990) Mathematical Methods in Linguistics. Kluwer Academic Publishers
• Petkevič, V. (1995) A New Formal Specification of Underlying Structure. Theoretical Linguistics, vol. 21, No.1
• Sgall, P., Hajičová, E., Panevová, J. (1986) The Meaning of the Sentence in Its Semantic and Pragmatic Aspects. D. Reidel Publishing Company, Dordrecht/Academia, Prague
• Štěpánek, J. (2006) Závislostní zachycení větné struktury v anotovaném syntaktickém korpusu. PhD Thesis, MFF UK
![Page 55: Prague Dependency Treebank - ÚFALufal.mff.cuni.cz/~lopatkova/2017/docs/1-intro-trees-cor.pdf · Prague Dependency Treebank: ... Outline of the lecture ... constructions like topicalization,](https://reader030.vdocument.in/reader030/viewer/2022020120/5b42247c7f8b9a997a8b4d1b/html5/thumbnails/55.jpg)
other non-dependency relations in PDT: • technical root – effective root of a sentence • syntactically unclear expressions rhematizers; sentence, linking and modal adverbial expressions, conjunction
modifiers
• list structures names, foreign expressions
• phrasemes
Dependency and non-dependency relations
přijede
otec asi.MOD zítra
#Forn
císař
čínský
Chou chun Tung
#Idph
číst
a #Gen
Timur parta
#PersPron
široko
daleko.DPHR
PDT – Intro Lopatková