Corpus Mark-up
UoL Summer Institute in Corpus
Linguistics
Matthew Brook O’Donnell
Aims
• Introduce the concepts of corpus
mark-up and annotation
• Consider why we would want to add
extra non-textual information to
corpus texts
• Use a pos-tagger and tagged text
What is Corpus Annotation?
• ‘the practice of adding interpretative
linguistic information to a corpus’
(Leech 2005)
– interpretative
– linguistic
– results in -> value-added corpus
Terminology
• Corpus Markup
– processing/formatting information
– metadata/text classifications
– structural representation
• Tagging
– (usually) inline addition of category to word(s)
• Parsing
– higher-level, multiword units (constituents)
– chunking/shallow vs. full syntactical parsing
– needn’t just be syntactical analysis
• XML
– eXtensible Markup Language
Why Annotate?
1.
2.
3.
4.
Manual examination of corpus
Automatic analysis of corpus
Reusability of annotations
Multi-functionality
Leech 2005
5. Objective record of analysis
McEnery 2003
6. Annotation process is corpus analysis
O’Donnell 1999
Types of Corpus Annotation
•
•
•
•
•
•
•
•
Part-of-speech (POS)
Lemmatization
Syntactical (parsing)
Semantic (domain classifications)
Coreference (Discourse)
Pragmatic (Speech acts – dialogue)
Stylistic
Research specific (ad hoc)
POS Tagging: Claws C5
Corpus_NN1 annotation_NN1 is_VBZ
the_AT0 practice_NN1 of_PRF
adding_VVG interpretative_AJ0
linguistic_AJ0 information_NN1
to_PRP a_AT0 corpus_NN1 ._.
NN1 singular noun
AJ0 adjective (unmarked)
VBZ -s form of the verb "BE“
PRF the preposition OF
VVG -ing form of lexical verb
AT0 article
POS Tagging: Claws C7
Corpus_NN1 annotation_NN1 is_VBZ
the_AT practice_NN1 of_IO
adding_VVG interpretative_JJ
linguistic_JJ information_NN1
to_II a_AT1 corpus_NN1 ._.
http://www.comp.lancs.ac.uk/ucrel/claws/trial.html
POS Tagging: POSTagger
Corpus/NN annotation/NN is/VBZ
the/DT practice/NN of/IN
adding/VBG interpretative/JJ
linguistic/JJ information/NN
to/TO a/DT corpus/NN ./.
Parsing: Chunking
[NP (NN Corpus) (NN annotation) ]
(VBZ is)
[NP (DT the) (NN practice) ]
(IN of) (VBG adding)
[NP (JJ interpretative) (JJ linguistic) (NN
information) ]
[PP (TO to) [NP (DT a) (NN corpus) ]
Parsing
(S
(NP Corpus annotation)
(VP is
(NP
(NP the practice)
(PP of
(S (VP adding
(NP interpretative linguistic
information)
(PP to (NP a corpus))
))
)
)
)
.)
Semantic Annotation
• Each word given code from
thesaurus-style dictionary
• Also called Word Sense Tagging
• Examples
– UCREL Semantic Analysis System
[http://www.comp.lancs.ac.uk/ucrel/usas/]
– WordNet
[http://wordnet.princeton.edu/]
Semantic Annotation
•
The noun move has 5 senses (first 5 from tagged texts)
•
1. (377) move -- (the act of deciding to do something; "he didn't make a
move to help"; "his first move was to hire a lawyer")
•
2. (70) move, relocation -- (the act of changing your residence or place of
business; "they say that three moves equal one fire")
•
3. (57) motion, movement, move, motility -- (a change of position that does
not entail a change of location; "the reflex motion of his eyebrows revealed
his surprise"; "movement is a sign of life"; "an impatient move of his hand";
"gastrointestinal motility")
•
4. (30) motion, movement, move -- (the act of changing location from one
place to another; "police controlled the motion of the crowd"; "the
movement of people from the farms to the cities"; "his move put him
directly in my path")
•
5. (5) move -- ((game) a player's turn to take some action permitted by the
rules of the game)
Semantic Annotation
•
The verb move has 16 senses (first 13 from tagged texts)
•
1. (130) travel, go, move, locomote -- (change location; move, travel, or
proceed; "How fast does your new car go?"; "We travelled from Rome to
Naples by bus"; "The policemen went from door to door looking for the
suspect"; "The soldiers moved towards the city in an attempt to take it
before night fell")
•
2. (60) move, displace -- (cause to move, both in a concrete and in an
abstract sense; "Move those boxes into the corner, please"; "I'm moving
my money to another bank"; "The director moved more responsibilities onto
his new assistant")
•
3. (52) move -- (move so as to change position, perform a nontranslational
motion; "He moved his hand slightly to the right")
•
4. (20) move -- (change residence, affiliation, or place of employment; "We
moved from Idaho to Nebraska"; "The basketball player moved from one
team to another")
Tools
• XML
• Annotation Editors
– GATE
• WordSmith
The ‘Great Annotation
Debate’
• Leech et al. ‘annotation = value added’
• Sinclair ‘annotation = perilous activity’
• Scott ‘beware of the POS prison!’
Sinclair on the perils of
corpus annotation
• ‘The interspersing of tags in a
language text is a perilous activity,
because the text thereby loses
integrity…’
‘Current Issues in Corpus Linguistics’ (Sinclair 2004: 191)
Sinclair on the perils of
corpus annotation
• ‘..one cosy consequence of using
tagged text is that the description
which produces the tags in the first
place is not challenged – it is
protected. The corpus data can only
be observed through the tags; that is
to say, anything the tags are not
sensitive to will be missed’
‘Current Issues in Corpus Linguistics’ (Sinclair 2004: 191)
Sinclair on the perils of
corpus annotation
• ‘In corpus-driven linguistics you do
not use pre-tagged text, but you
process the raw text directly and
then patterns of this uncontaminated
text are able to be observed.’
‘Current Issues in Corpus Linguistics’ (Sinclair 2004: 191)
Hunston – annotation as
‘double-edged sword’
• ‘…the categories used to annotate a
corpus are typically determined
before any corpus analysis is carried
out, which in turn tends to limit, not
the kind of question that can be
asked, but the kind of question that
usually is asked.’
(Hunston 2002: 93)
Hunston – annotation as
‘double-edged sword’
• ‘Most of the work that is done using
annotated corpora uses categories that
have been developed in pre-corpus days,
such as nominal clauses, anaphoric
reference… Phenomena such as frames
or semantic prosody… tend to have been
identified from plain text corpora and
word-based studies.’
(Hunston 2002: 93)
Corpus-based approach
ANALYSIS
categorization
plain corpus
annotated
corpus
CORPUS
METHODS
ANALYSIS
Annotate
Corpus
• POS
• Parsing
• Semantic
• Reference
generalization
RESULTS
DATA
Corpus-driven approach
plain corpus
CORPUS
METHODS
DATA
ANALYSIS
generalization &
categorization
RESULTS
Problem for both CB & CD
Approach
• Serial/Sequential process
– CB analysis before (annotation) and
after processing
– CD analysis only after processing (so no
need for annotation)
• Empirical process is cyclic
– analysis feeds back into process and
around again… and again…
So what if….
• Hunston - ‘Most of the work that is done using
annotated corpora uses categories that have
been developed in pre-corpus days….’
(Hunston 2002: 93)
• we annotate categories that have come
out of corpus analysis instead of/as well
as traditional categories?
New uses for corpus
annotation
•
Cyclic investigation process
–
How sould we annotate:
1. KWIC/Frequency list/Collocates etc.
2. Annotate results
3. Goto 1
–
–
–
–
collocates
lexical items
semantic associations/prosodies
Local textual functions
References
Leech, G
2005 ‘Adding Linguistic Annotation’, in M. Wynne,
Developing Linguistic Corpora: a Guide to Good Practice
(Oxford: Oxbrow Books), pp. 17-29
[http://ahds.ac.uk/linguistic-corpora/]
Hunston, S.
2002 Corpora in Applied Linguistics (Cambridge: Cambridge
University Press)
McEnery, A
2003 ‘Corpus Linguistics’, in R. Mitov (ed.), The Oxford
Handbook of Computational Linguistics (Oxford: Oxford
University Press), pp. 448-463
References
O’Donnell, M.B.
‘The Use of Annotated Corpora for New Testament
Discourse Analysis: A Survey of Current Practice and
Future Prospects’, in S.E. Porter and J.T. Reed (eds.),
Discourse Analysis and the New Testament: Results and
Applications (Sheffield: Sheffield Academic Press, 1999),
pp. 71-117.
Sinclair, J.
2004 Trust the Text: Language, Corpus and Discourse
(London: Routledge)