1. Overview of Corpus Linguistics
10/8/14
Overview of Corpus Linguistics
Ling 240
© 2007 Microsoft Corporation. All rights reserved. Microsoft, Windows, Windows Vista and other product names are or may be registered trademarks and/or trademarks in the U.S. and/or other countries.
The information herein is for informational purposes only and represents the current view of Microsoft Corporation as of the date of this presentation. Because Microsoft must respond to changing market conditions, it should not be interpreted to be a commitment on the part of Microsoft, and Microsoft cannot guarantee the accuracy of any information provided after the date of this presentation.
MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION. 1

2. Outline
Definition
History
Current status

3. What is corpus linguistics?
Linguistics: the scientific study of language
using…
Corpus: a large and principled collection of natural texts

4. History of corpus linguistics
As early as 1897, Wilhelm Kaeding compiled a had 5,000 people compile a corpus of 11 million German words (and calculate their frequency, distribution of letters).
In the early 1900s, Otto Jesperson, a Danish professor filled shoeboxes with thousands of paper slips containing interesting English sentences.
In 1959, Randolph Quirk started the Survey of English Usage of spoken and written language which he used to create a comprehensive English grammar.

5. History of corpus linguistics
1961: Brown Corpus
1M words
500 samples of 2,000 words
Various genres; printed, edited, American English
1961: Lancaster-Oslo/Bergen (LOB) Corpus
British version of Brown Corpus
1991: Frown and FLOB Corpora
1988: International Corpus of English (ICE)
World English varieties
20 completed so far

6. History of corpus linguistics
1991: British National Corpus (BNC)
100M words
Wide range of written (90%) and spoken (10%) texts
2008: BYU Corpora
Corpus of Contemporary American English (COCA)
TIME corpus
Corpus of Historical American English (COHA)
GloWbE Corpus
International Corpus of Learner English (ICLE)
MICASE & MICUSP

7. Status of corpus linguistics
Is corpus linguistics a branch of linguistics or a method for doing linguistics?
Evidence for branch:
Journals such as Corpora and the International Journal of Corpus Linguistics
Some researchers claim corpus linguistics as their area of emphasis
Evidence for method:
Most linguistic phenomena can be measured using CL
CL has the potential to inform virtually any theory

8. Characteristics of corpus-based analyses
It is empirical, analyzing the actual patterns of use in natural texts;
It utilizes a large and principled collection of natural texts, known as a “corpus” as the basis for analysis;
It makes extensive use of computers for analysis, using both automatic and interactive techniques;
It depends on both quantitative and qualitative analytical techniques

9. Uses of corpora Changes over time Changes in register
Changes in situation
Changes in individual

10. Time

11. Different Genitives Of genitive The leg of the table 's genitive
The table's leg
NN genitive
The table leg

12. ‘s genitive vs. of-genitive vs. NN sequence

13. NN sequence across time in three registers

14. Situation

15. Phrasal Compression Uncompressed Compressed
The dog that was hungry was looking for something to eat.
Drugs that require a prescription should be monitored
Compressed
The hungry dog was looking for something to eat.
Prescription drugs should be monitored.

16. Phrasal compression across levels in an EAP reading series

17. Phrasal compression across levels in another EAP reading series

18. Individual

19. ‘Abstract Exposition versus Concrete Action’

20. Corpus Design and Representativeness
Ling 240

21. Designing Representative Corpora
Many people believe that the design of a corpus doesn’t matter as long as it is large enough.
Researchers typically focus on target domain representativeness and ignore linguistic representativeness
Target domain (medical texts, newspapers, academic, general English, spoken)
Very few corpora are actually evaluated in terms of their representativeness (target domain or linguistic)

22. Steps—representing the target domain
Describe the target domain
Design the corpus to represent target domain
Complete the sampling
Simple random
Randomly choose sections of the data for the corpus
Stratified
Determine what genres are included and randomly sample from those data
Cluster
Divide data into naturally occurring groups and sample from them

23. Norming practice (raw count/total words) * 1000 Text A Text B # Nouns50
100
# Words
200
1000
(raw count/total words) * 1000
Text A: (50 nouns / 200 words) * 1000 = 250 nouns per thousand words
Text B: (100 nouns / 500 words) * 1000 = 200 nouns per thousand words

24. Norming practice BNC has 100 million words COCA has 450 million words#
Per M
snuck 11
767
sneaked
132
830

25. Corpus Annotation
Ling 240

26. Annotation
Corpora can be annotated for a wide range of external and internal variables.
External variables
Speaker
L1 background
Gender
Extralinguistic information (e.g., laughter, nodding, etc.)

27. External annotation—example
<Exam ID: 3B>
<Arrangement ID: 54945>
<Center ID: 14>
<Candidate ID: 42285>
<Test Date: 12/6/2013>
<Age: 19>
<Gender: F>
<L1: Arabic>
<Reason for test: B>
<Original MELAB: 2>
<Original Transformed: 3>
<Second MELAB: >
<Second Transformed: >
<End header>
E: Alright, welcome to the MELAB speaking exam, my name is <deleted>. And uh what is your name?
T: Uh my name is uh <deleted>.
E: Now I'll just uh read the MELAB ID number that we have for you. Uh you don't need to know it or anything. The number is <deleted>. Alright now that's out of the way. Uh why don't you uh start by telling me a little bit about why you're taking the MELAB today.
T: Uh actually I came to USA to complete my education here, so uh if I want to go to university I need to get score and to to be good in speak English and I take uh a lab exam so I can enter to the university.
E: Okay uh so uh what are you interested in studying at the university?
T: Actually I think about uh medical science.

28. Part of speech tagging Rule-based Probability-based 95%+ accuracy rate
Some features very easy (e.g., the)
Some features more difficult (e.g., that)
Pronoun (He doesn’t like that.)
Determiner (He doesn’t like that dog.)
Complementizer (They thought that they could do it.)
Relativizer (The thought that I entertained.)

29. POS tagging accuracy Accuracy Example
Precision – What percent of the cases labeled as X are actually X?
Recall – What percent of all of the true cases of X were labeled as X?
Example
He saw that dog that I saw.
If both ‘that’s are tagged as determiners:
Calculate the precision and recall for determiners
Calculate the precision and recall for relativizers

30. POS tagging—two examples
CLAWS Tagger
Biber Tagger
Everything_PN1
I_PPIS1
've_VH0
read_VVN
says_VV0
they_PPHS2
were_VBDR
warned_VVN
to_TO
leave_VVI
immediately_RR
Everything ^pn++++=Everything
I ^pp1a+pp1+++=I've
've ^vb+hv+aux++0=EXTRAWORD
read ^vprf+++xvbnx+=read
says ^vb+vpub+++=say's
they ^pp3a+pp3+++=they
were ^vbd+bed+aux++=were
warned ^vpsv++agls+xvbnx+=warned
to ^to+vcmp+++=to
leave ^vbi++++=leave
immediately ^rb+tm+++=immediately

31. Lemmatization Lemma The citation or dictionary entry Run is the lemma
It includes the words run, running, runs, ran
We often want the frequency of the lemma not of a particular word like running

32. Answer these questions about COCA
What external annotation does it contain?
What internal annotation does it contain?

33. Answer these questions about COCA
What external annotation does it contain?
Text source
Date of publication
What internal annotation does it contain?
Lemmatization
Part of speech
Genre

34. Example: ‘s-’ versus ‘of-genitive’
‘the bird’s owner’ vs. ‘the owner of the bird’
Finding 1: “by 1991, the s-genitive had overtaken the of-genitive in frequency” (Leech, et al., 2009)
Finding 2: of-genitive is almost 10 times more frequent than the s-genitive in present-day English (Longman Grammar)
Q: Are these findings contradictory??? 34

35. Corpus study design—variationist
Two approaches to corpus linguistics: Variationist and Text-Linguistic (Biber, 2012)
Variationist: “has the goal of comparing linguistic variants: whether one or the other variant is preferred,” and identifying factors that predict which variant is used (Biber, 2012).
Statistics: Binomial/logistic regression; Linear discriminant analysis
Interpretation: When a choice can be made, variant X is preferred over variant Y, and factors A, B, and C play a role. 35

36. Variationist Analysis (Type A)
Unit of analysis is linguistic feature
Most studies do not take register into account (e.g. collocational studies)
Comparison of the proportion of use in a particular register
E.g., Benedict Szmrecsanyi & Hinrichs, 2008
preference of s-genitive over of-genitive in speech; BUT: s-genitives overall more frequent in writing.

37. Corpus study design—text-linguistic
Text-Linguistic: “has the goal of providing a linguistic description of texts, by describing the density of grammatical features in texts” (Biber, 2012)
Statistics: T-test, ANOVA, Multiple regression, Factor analysis
Interpretation: Feature X is more frequent in context A than context B; or Feature X is more frequent than feature Y 37

38. Text-linguistic (Type B)
Comparison of actual frequency of use in a particular register
Unit of analysis is text
Normed rates of occurrence by text
Much more common for register studies

39. Text-linguistic (Type C)
Also compares frequencies of use in a particular register
Unit of analysis is subcorpus
Normed rates of occurrence for features across subcorpora
Cannot use inferential statistics (need to look at individual text to get a mean score for the register)

40. Quantitative analysis
Coding/tagging features
Counts in text vs. subcorpus
Norming (raw count/total words * 1000)
Use appropriate statistical tests if applicable

41. Kinds of Corpora Spoken language General corpora (mainly written)
Bitext (two languages side-by-side)
Specialized
Children’s speech
L2 learner speech
Historical

42. General Corpora Mainly written British National Corpus (BNC)
100 million words
10% spoken
25% fiction
75% non-fiction

43. General Corpora Corpus of Contemporary American English (COCA)
450 million words (more added every year)
Divided into registers
Spoken
Fiction
Academic

44. General Corpora International Corpus of English
1 million words from each country
60% spoken, 40% written

45. Historical Corpora Helsinki Corpus English texts from 770-1700
Corpus of Historical American English (COHA)
1860-present

46. Introduction to COHA Corpus of Historical American English
End up verbing
Try and verb versus try to verb
Adjectives and nouns used in 2000s not before
Collocates of Muslim, liberal, Mormon ?

47. Raw Corpora Not easily searchable Not tagged Project Gutenberg
Pre 1928 books (copyright expired)
Online newspapers
Time Magazine
The internet
General Conference

48. Where can you get corpora?
Online
BNC, COCA, COHA
Distributors (membership required)
ELRA (based in Europe)
Linguistic Data Consortium
US based
BYU has a membership
Catalog
Top 10 corpora