Qiufang Wen

1
Chinese learner corpora and second language
research
The 2006 International Symposium of
Computer-Assisted Language Learning June 2-4,
2006, Beijing

• Qiufang Wen
• The national research center for foreign language
  education, BFSU

2
Topics to be addressed

• English corpora of Chinese learners
• Corpus-based studies on English learners in
  mainland China
• Several corpus-based studies on English learners
  interlanguage by myself or together with my
  colleauges
• Advantages and disadvantages of corpus-based
  studies on the interlanguage

3
Topic One

• English corpora of Chinese learners

4

• Chinese learner English Corpus (CLEC)
• College Learners Spoken English Corpus (COLSEC)
• Spoken and Written Corpus of Chinese Learners
  (SWECCL)
• Version 1
• Version 2 (under construction)
• Bilingual Corpus of Chinese English Learners
  (BICCEL) under construction

5
1. Chinese learner English Corpus (CLEC) by Gui
Yang in 2003

• Written corpus 1 million
• Timed and untimed compositions
• Levels of proficiency
• Middle school students
• Non-English major (Band 4)
• Non-English major (Band 6)
• English majors (Band 4 )
• English majors (Band 8)
• Error-tagged

6
Two Types of English Learners in University

• English Majors Non-English majors

Year 4 Year 3 Year 2 Year 1
Year 4 Year 3 Year 2 Year 1
Band 8
Band 6
Band 4
Band 4
Band 2
7
2. College Learners Spoken English Corpus
(COLSEC) by Yang Wei in 2005

• Tokens 0.7million
• Source National spoken English test for
  non-English majors
• Test items
• Teacher-student conversation
• Student-student discussion
• teacher-student discussion
• Data format written transcripts

8
3. Spoken and Written Corpus of Chinese Learners
(SWECCL) by Wen, Wang Liang in 2005 (Version 1)

SWECCL
WECCL
SECCL
1.18 million
1.46 million
9
Spoken (SECCL)

• Source of data
• National spoken English test 1996-2002
• Second-year English majors
• Data format
• Digital sounds as well as transcripts of the
  speeches

10
National spoken English test for English majors
Band 4

• Test format
• Test in a lab
• The number of testees annually
• 2006 more than 16,000
• Expect to have 50,000 in the future
• Scoring procedures
• A random sample (30-35 tapes)
• Two raters scoring one tape independently

11

• Number of subjects
• 6 groups from each year (1996-2002)
• 42 groups (30/35) about 1400 students
• About 230 hourss speech
• Testing items

12
Testing items
13
The structure of SECCL
Tagged
Article
Past Tense
Text
Special
Whole
Task A
SECCL
Raw
Task
Task B
Task C
Year
Sound files
(1996-2002)
14
The written component
Written
Year 1
Year 2
Year 3
Year 4
15
The written component

• Source of data
• Timed compositions in class (40 minutes, no less
  than 300 words)
• Take-home compositions (no word limit)
• Types of compositions
• Argumentative (a list of topics provided)
• Narrative

16
SWECCL in 2007 (Version 2)

SWECCL
WECCL
SECCL
Two million
Two million
17
SECCL(Version 2)

• 2003-2006 National Spoken English Test for
  second-year English majors (band 4)
• 2000-2006 National Spoken English Test for
  4th-year English majors-Band 8 (Task 3)
• A longitudinal data (2001-2004)

18
Spoken (Band 8)

• Testing item (Task C)
• Make a comment on a given topic
• Data format
• Digital sounds as well as transcripts of the
  speeches

19
Spoken (Longitudinal)

• 72 students 56 students
• 40 hours speech

20
Tasks

• Reading aloud
• Retelling a story
• Talking on a given topic (Narrative)
• Talking on a given topic (argumentative)
• Conversation (Role play)
• Discussion on a given topic

21
4. Bilingual Corpus of Chinese English Learners
(BICCEL)
BICCEL
Spoken
Written
E-C
C-E
E-C
C-E
0.5 million
0.5 million
0.5 million
0.5 million
22
Spoken component of BICCEL

• National Oral English test Band 8
• The 4th year English majors
• Interpreting from English to Chinese (Task A)
• Interpreting from Chinese to English (Task B)
• 2001-2005 1100 testees

23
Written component of BICCEL

• Source of data in-class assignment
• E-C and C-E translation
• Across the 3rd and 4th years
• 30 universities across the country

24
Topic Two

• A brief review of corpus-based studies on Chinese
  learner English

25
Sources

• China National Knowledge Infrastructure
  (CNKI)(On-line journals)
• Digital dissertation database

26
Corpus-based studies in mainland China
27
Research areas
28
Conferences workshop

• The International conference on Corpus
  Linguistics 25-27 October, 2003
• The First National Symposium on corpus
  linguistics and ELT Education
• 11-13 October, 2004
• Workshop on the use of corpus in teaching and
  research 17-19 March, 2006

29
Topic Three

• Several corpus-based studies on English learners
  interlanguage by myself or together with my
  colleagues

30
Study One

• Features of oral style in English compositions of
  advanced Chinese EFL learners
• (Wen, Q.F. Ding, Y.R. Wang, W.Y. 2003, Foreign
  Language Teaching Research (4)268-274.

31
Study Two

• A Study on Frequency Adverbs Used by Advance
  English Learners in China
• Wen, Q. F. Ding, Y. R. 2004. Modern foreign
  languages(2) 141-147.

32
Study Three

• An analysis of English Majors Abstracting
  abilities through their English compositions
• Wen, Q.F. Liu, R.Q. 2006. Foreign Languages
  (2)

33
Study Four

• A longitudinal study on the developmental
  features of speaking vocabulary by English majors
  in mainland China
• Wen, Q. F. 2006. Foreign Language Teaching
  and Research (3).

34
Study Five

• A comparison of developmental features of
  Speaking and Writing vocabulary by English majors
  
• Wen, Q. F. 2006. Foreign languages and Foreign
  Language Teaching (4)

35
Study Six

• Patterns of change in speaking vocabulary
  development by English majors

36
Study Two

• A Study on Frequency Adverbs Used by Advance
  English Learners in China
• Wen, Q. F. Ding, Y. R. 2004. Modern foreign
  languages(2) 141-147.

37
Frequency Adverbs

• Adverbs used for describing how often something
  happens
• never, sometimes, usually, always

38
Top Twenty Frequency Adverbs

• Most frequently used by native
• speakers according to the analyses of the
  British National Corpus (BNC) by Leech, Rayson
  and Wilson (2001)

39
Top Twenty Frequency Adverbs (TTFAs)
40
Common features

• All high-frequency words
• Different frequencies in speech and writing
  except sometimes and twice (Leech et al. 2001)

41
A comparison of TTFAs in speech and writing

• The overall difference
• TTFAs more likely occur in writing than in
  speech.
• The specific differences
• Speech never, always, ever, normally
• Neutral sometimes, twice
• Writing 14 words

42
Previous corpus-based studies

• e.g. Altenberg Granger, 2001 Cobb, 2002
  Ringbom, 1998 Wen, Ting, Wang,2003
• Conflicting finding one overuse vs. underuse

43
Examples

• Overuse high-frequency words in writing (Cobb,
  2001)
• Overuse modal verbs (Aijmer, 2002)
• Underuse adverbial connectors (Altenberg
  Tapper, 1998)
• No study on frequency adverbs

44
Conflicting finding two

• Tend to use written style features in their
  speech
• Tend to use a mixed register in either speech or
  in writing
• Tend to use oral style features in their writing
• Did not compare the use of high-frequency words
  in speech with writing

45
General purposes of this study

• Whether Chinese EFL learners simply overuse the
  TTFAs or they overuse some while underusing
  others
• whether they use the TTFAs similarly or
  differently when compared their speech with
  writing

46
Research questions

• Do they overuse or underuse the TTFAs differently
  between speech and writing?
• Do they differ more from native speakers in
  writing or in speaking with regard to the use of
  the TTFAs?
• Do they demonstrate a similar pattern of
  writing-speaking difference as native speakers in
  the use of the TTFAs?

47
Data for analysis
48
Data analysis

• Four comparisons
• Learners speech and native speakers speech
• SECCL vs. BNCS
• Learners writing and native speakers writing
  CLEC vs. BNCW
• Dif. in learners speech native speakers and
  Dif. In learners writing native speakers
• SECCL vs. BNCS and CLEC vs. BNCW
• Dif. In learners speech writing and dif. in
  native speakers speech writing
• SECCL vs. CLEC and BNCS vs. BNCW

49
Results(1)

• TTFA use in learners spoken corpus (SECCL)

50
Results(2)

• TTFAs use in learners written corpus(CLEC)

51
Results(3)

• Comparison of learners speech with their
  writing in TTFA use (Overuse)

52
Results(3)

• Comparison (Underuse)

53
Results(3)

• Comparison (identical or similar)

54
Results(4)

• Speaking-writing differences in TTFA use in the
  CEMIC and the BNC

55
Results(4)

• Speaking-writing differences in TTFA use in the
  CEMIC and the BNC

56
Summary (1)

• English majors in China tend to overuse and
  underuse certain TTFAs in their speech and
  writing. The overuse tendency is stronger than
  the underuse tendency in both speech and writing.

57
Summary (2)

• The overuse tendency is more marked in their
  speech than in their writing while the underuse
  tendency is also slightly stronger in speech than
  in writing. Some of the overused or underused
  TTFAs in speech are the same as those in writing
  but others are different.

58
Summary (3)

• Chinese English majors demonstrate a pattern of
  speaking-writing difference that is opposite to
  that shown in the native speakers corpus they
  tend to use more TTFAs in their speech than in
  their writing while native speakers tend to use
  more TTFAs in their writing than in their speech.
  This shows that Chinese EFL learners use TTFAs
  without awareness of their register differences.

59
Possible reasons

• Limited vocabulary (Table 1b)
• Use them as time buyers
• Without equivalents readily available in Chinese

60
Topic Four

• Advantages and disadvantages of corpus-based
  studies on SLA

61
Advantage One

• A large sample stored electronically and open to
  the public
• Validity and reliability (replicable)
• Possible for a diachronic study

62
Advantage Two

• Using a computer software such as WordSmith
• Effectiveness and efficiency

63
Advantage Three

• Understand the learner language from a different
  perspective
• Correct vs. incorrect
• More acceptable vs. less acceptable
• Frequency
• Overuse
• Underuse
• unuse

64
Disadvantages
65
Closing Remark

• The number of researchers increasing
• Constructing different types of corpora
• Carrying corpus-based studies
• Findings useful for textbook writers as well as
  for practitioners

66

• Thank you!!!