The Pedagogic Value of Corpora:

1

• The Pedagogic Value of Corpora
• A Critical Evaluation
• Lynne Flowerdew

2
Key issues in corpus-based pedagogy

• Five key issues
• Discovery learning vs. incidentalism
• Inductive vs. deductive approaches
• Corpora and the use of other resources
• Context and co-text
• The student as linguistic researcher
• Issues examined with reference to
• http//Mywords.ust.hk/ (suite of programs)
• http//ysomeya.hp.infoseek.co.jp/ (1
  milllion-word corpus of business letters)

3
Phraseological approach to language analysis

• Probabilistic, syntagmatic view of language
• Collocation
• Colligation
• Semantic prosody
• Semantic preference

4
Examples of Phraseology

• Sinclairs (1999 19-20) analysis of the 31
  instances of the item budge in a corpus of 20
  million words
• 25 of 31 instances of budge are intransitive
• ght be out of his mind and refuse to
  budge. In that case the Vice-Pres
• sat in a corner, I determined not to budge
  from it until closing time. I
• tried the idea on him. He wouldnt budge. He
  seemed to have already
• item that does not budge is the subject.
• where item is animate, it strongly collocates
  with refuse.
• semantic preference of refusal is found in most
  other instances, mainly through colligation with
  certain modals.

5
Examples of Phraseology (contd.)

• In the transitive instances of budge
• hesitated. He knew he couldnt budge Ben Canaan.
  He walked to the alc
• but she knew she could not budge me from my
  view. We spent severa
• the non-budging item is object.
• the semantic preference is inability with strong
  colligation with modals of ability.
• from the perspective of the person who wants
  something moved, the semantic prosody is negative
  associated with feelings of frustration and
  irritation.

6
Issue 1

• Discovery learning vs. incidentalism

7
Viewpoints

• The Learner as Traveller (Bernardini 2002,
  2004)
• Swales comments that this approach, and the field
  of corpus linguistics as a whole, is marked by a
  kind of incidentalism (Swales 2002, Lee Swales
  2006)

8
Example 1

• Student query
• This project focuses on the incidence of
  mosquitoes on campus.
• This project is focused on the incidence of
  mosquitoes on campus.

9
Word Neighbors (Milton 2004, 2006)
10
Search in Reports subcorpus
11
Triggered Query

• Why are there so many occurrences of focus in
  the present perfect?

12
Click on Show results
13
Example 2

• Student query
• We plan to do a survey on the usage of computers.
• We plan to carry out a survey on the usage of
  computers.

14
Search in Reports subcorpus
15
Link to Just The Word
16
Search in Just The Word
17
Click on do survey
18
Triggered query

• Can we use implement with survey?

19
Search in Just the Word
20
Triggered query

• What verb do we use with survey questionnaire?

21
Search in Just the Word
22
Remarks

• This type of incidentalist learning may
• be very time-consuming
• involve a lot of trial-and-error
• put the teacher on the spot
• be too unsystematic for some students

23
Remarks (contd.)

• But, discovery-based learning through browsing
  the corpus
• encourages noticing (Swain 1998)
• encourages students to examine language at a
  discourse level (Granger 1999 Hahn 2000)
• encourages students to attain a more autonomous,
  divergent approach to learning
• can be motivating real-time processing of
  queries

24
Issue 2

• Inductive vs. deductive approaches

25
Inductive starting point

• Student query
• With a very crowded schedule, students level of
  motivation was decreased.
• With a very crowded schedule, students level of
  motivation has decreased.
• Corpora useful for phraseological queries.

26
Change-of-state verbs three voices

• Overpassivisation of ergative verbs.
  Celce-Murcia (2002 146) notes
• With the verbs increase and decrease the
  ergative tends to be used when the inanimate
  subject is objectively or subjectively measurable
  (rather than an animate agent/dynamic instrument
  object both of which favor active voice or a
  patient subject for the passive voice.

27
Search in Word Neighbors
28
Pedagogic processing of corpus data

• Students sometimes need to be given prompts to
  work out the tendencies of phraseological
  patterns, e.g.
• Do you notice any difference in the subjects for
  was decreased and has decreased?

29
Deductive starting point

• Student queries
• A good user-friendly interface can reduce the
  time to familiarise the program.
• A good user-friendly interface can reduce the
  time to familiar with the program.
• It may be time-consuming for students to search
  the corpus data to extrapolate the rule.

30
Clues for an item (My Words, Milton 2004)
31
From grammar guide to
32
corpus examples
33
Remarks

• Whether an inductive or deductive approach is
  used depends on
• Nature of the query
• grammar rule
• phraseological tendency
• Clues and prompts can be used to mediate the
• inductive deductive
  continuum.
• Cognitive style of student (field dependent vs.
  field independent)
• Level and motivation of student

34
Dynamic paradigm for corpus investigations

• Inductive Phraseology
• (probabilities)
• (clues)
• Deductive Grammar rules

35
Issue 3

• Corpus resources vs. other resources

36
Which online resources?

• Chambers Sulllivan (2004 158)
• The concept of literacy now includes not only
  the knowledge and skills which are traditionally
  associated with that concept, but also the
  ability to select, evaluate and use the
  electronic tools and resources appropriate for
  the activity which is being undertaken.

37
Example 1

• Student query
• The mosquitoes around the university halls are
  irritating. These irritating insects cause
  students to feel uncomfortable.
• One student suggested searching in the whole
  Word Neighbours corpus.
• Another student suggested replacing irritating
  by ?????? insects.

38
Links to other online resources
39
Search in Wordsmyth Dictionary - Thesaurus
40
Search in Wordsmyth Dictionary Thesaurus
(contd.)
41
Google search

• annoying insects 24,600 hits
• vexing insects 5 hits
• Google search shows vexing collocates with
  problems, issues, matter, questions

42
Example 2

• Student query
• The Victoria Harbour is becoming more and more
  narrow.

43
Which sub-corpus?
44
Noticing of the data
45
Scrutiny and Analysis of Data
46
Issue 4

• Context and Co-text

47
Context

• Widdowson states that corpus data are
  decontextualised The data travels but the
  context does not travel with it.
• Data have to be transformed from samples of
  language to authentic examples to fit the
  students context of writing (Widdowson 2004).
• Pedagogic processing is necessary

48
Example 1

• Student writing
• Implementing our proposed changes is also
  highly welcome for students and outside firms and
  we can have a more effective curriculum preparing
  students for their future.

49
Search in Just The Word
50
Click on implementation would
51
Scrutiny and Analysis of Data

• Interpretation of corpus data
• Implementation result in or lead to
• negative semantic prosody
• Implementation provide or ensure
• positive semantic prosody
• Choice of verb to collocate with effective
  curriculum

52
Authentication and Reformulation

• Reformulation of corpus data for authentication
  (pragmalinguistic appropriacy)
• We suggest that implementation of our proposed
  changes would ensure a more effective curriculum
  to prepare students for their future.

53
Example 2 Business Letters

• Student writing
• I would appreciate if you can contact me
• I would be very appreciated if you can

54
appreciate it if.. (105 hits)
55
Itappreciated if (9 hits)
56
Absence of Context

• Concordance data provides formal features for
  self-editing.
• Concordance output reveals several variations for
  basic frames of both phrases.
• Concordance data is the product of communication
  and does not reflect the processes of
  communication (Aston 1995).

57
Appropriation of corpus data

• Which sample do students appropriate for their
  own writing?
• Pedagogic processing necessary for
  recontextualisation of data

58
Co-textual Environment
59
Pedagogic processing of data

• Carter McCarthy (1995) propose 3 Is.
• Illustration (looking at data)
• Interaction (discussion and sharing opinions and
  observations)
• Induction (making ones own rule for a particular
  feature)
• Elaboration of approach 4 Is.
• Illustration, Intervention, Interaction,
  Induction
• Intervention by teacher (hints)
• Co-text may provide some clues
• Frequency may be an indicator of usage

60
Issue 5

• To what extent should students be trained as
  linguistic researchers?

61
Training in Tools

• Römer (2006 127) on future developments
• With respect to concordance analysis packages,
  we have to consider whether the tools that are
  currently available are easy enough to use for
  learners and teachers who, obviously, have not
  had the same training as the corpus-linguistic
  researcher. Another item on my resources wish
  list would therefore be an easy-to-use, appealing
  concordance program (or a suite of programs) that
  teachers and learners would be willing and able
  to work with. it would be very useful if, with
  just one or two mouse clicks, some kind of
  advanced structuring of concordances (more that
  just an alphabetical sorting of the context)
  could be provided.

62
Example

• Student query
• There will be a great improvement on students
  attitude.

63
One click
64
Two clicks
65
Triggered query

• When do we use improvement over?

66
Three clicks
67
Students as corpus linguists?

• Johns (1991 2, cited in Gavioli 2005) mentions
  that effective language learning is itself a
  form of linguistic research.
• Students should be attuned to phraseological
  nature of language and encouraged to look at
  extended units of meaning
• Pedagogic processing of corpus data through
  clues and authenticating the data
• Students should be encouraged to browse the
  corpus set off in a certain direction, indulge
  in detours, observe the landscape, or press on
  until they find something noteworthy (Bernardini
  2002)

68
Conclusion Pedagogic value of corpora

• Corpora are very useful, but are one online
  resource among several.
• Future developments
• teachers
• students
  software
• library of
  developers
• online resources

69
Conclusion Pedagogic value of corpora

• Corpora are useful for
• Examining phraseologies
• Promoting a discovery-based, inductive approach
  to learning
• But,
• Learners need to acquire strategies for reading
  and interpretation (extended units of meaning)
• Pedagogic processing of corpus data

70

• Thank You!