Learner Corpus Research in EAP: Past, Present and Future

1
Learner Corpus Research in EAP Past, Present and
Future

• Lynne Flowerdew

2
Outline of talk

• Research in
• English for General Academic Purposes (EGAP)
• English for Specific Academic Purposes (ESAP)
• Some core issues
• Current and future directions
• Focus on academic writing

3
EGAP learner corpus research key areas

• Epistemic modality/stance markers
• Tense and aspect
• Vocabulary, collocations and phraseology
• (also connectors)

4
ESAP Learner Corpus Research
5
ESAP learner corpus research key areas

• Lexical bundles
• Keywords to phraseology
• Metadiscoursal expressions

6
Some core issues

• Starting point for analysis
• derived from existing framework or inductive
• corpus-based vs. corpus-driven (Tognini
  Bonelli)
• bottom-up vs. top-down
• Choice of reference/control corpus
• Explanation for data (taking into account
  variables)

7
EGAP Learner Corpus Research
8
Epistemic modality/stance markers

• Argumentative essays, e.g. ICLE sub-corpora
• governed by epistemic modality
• overlaps with Hylands (1999) stance markers
• hedges, boosters, self-mentions, attitude
  markers
• difficult for students to master complex
  interplay between hedges and boosters

9
1. Learner Corpus Research (L1 L2)

• Doubt and certainty (Hyland Milton 1997)
• 2 corpora, 500,000 words each
• Learner Native-speaker
• 150 essays by A level
  scripts
• HK students for British
  school leavers
• A Level UofEng of similar age
  education

10
1.Learner Corpus Research

• Procedure
• Inventory of 75 of most frequently occurring
  epistemic lexical devices from
• Holmes (1983) analysis of learned sections of
  Brown LOB
• Research literature on modality (Coates 1983)
• Reference grammars (Quirk et al. 1972)
• 50 sentences of each item(cut-off point, cf.
  Altenberg)

11
1. Learner Corpus Research

• Key findings
• Chinese learners limited range of devices,
  inappropriate strong convictions
• e.g. As I know, I am quite sure some parents are
  willing to pay vs.
  native-speaker
• e.g. On balance, it would seem that the only
  real solution to the problem would be to.
• Explanation
• due to lack of awareness of socio-pragmatic norms
• Teaching induced effect tutorial schools
  

12
2. Learner Corpus Research (L1 L2)

• Impact of culture on use of stance exponents in
  GRICLE (Hatzitheodorou Mattheoudakis 2011)
• 2 sub-corpora of around 200,000 words each
• Learner
  Native-speaker
• Greek component
  LOCNESS (not all timed)
• of ICLE (timed)
  PELCRA (timed)
• 
  (cf. Ädel 2008)
• Procedure
• Focus on adverbials (hedges, boosters, attitude
  markers)
• Hylands (2005) categories of metadiscourse as
• reference point
• own list (e.g. happily)

13
2. Learner Corpus Research (L1 L2)

• Results
• Motivation for research initial reading (Greek
  learners preference for wide range of boosters)
• Preliminary observations borne out by results
  Greek learners more emphatic extensive use of
  boosters
• Explanation (with reference to cultural factors)
• Consulted 1.7 million word Hellenic National
  Corpus
• Culturally induced from the Greek, authoritative
  style of writing
• NB unlike Hyland Miltons (1997) study and
  GRICLE (2011)
• study, McEnery Kifle (2002) found overuse of
  hedging in
• argumentative writing of Eritrean students

14
3. Learner Corpus Research (L2 vs. L2)

• Use of modal and reporting verbs in expression of
  stance (Neff et al. 2003)
• Corpora
• 5 learner sub- corpora from ICLE (French,
  Spanish,
• Italian, Dutch, German)
• 290,000 words in French . 195,000 in Spanish
• LOCNESS (150,000 words) as native reference
  corpus
• Focus on
• modal verbs (can, could, may, might, must)
• 9 reporting verbs (suggest, argue etc. among 12
  most frequently used in preliminary analysis)

15
3. Learner Corpus Research (L2 vs. L2)

• Results explanations
• can overused by all groups of non-native writers
  (massive overuse in Italian and Spanish
  sub-corpora)
• L1 transfer involving meaning, typology and
  sociolinguistic norms
• Typology
• Se puede apreciar un contraste entre...
• It can be appreciated a contrast between
• Positive politeness strategies
• The problems that we can find
• Developmental factors may play a role
• (cf. Aijmer 2002 Gilquin Paquot 2008)

16
4. ESAP Learner Corpus Study

• Stance options in data-description task in
  statistics (Wharton 2012)
• Corpus 40 student texts (4705 words)
• Procedures
• Inductive using Nvivo
• (small set of data in under-researched genre)
• Examination and reexamination of texts
• Common Content Assertions (5 propositions)
• Common Stances in Assertions (bare, hedged,
  vague, boosted, reader-inclusive)

17
ESAP Learner Corpus Research
18
ESAP learner corpus research key areas

• Lexical bundles
• Keywords to phraseology
• Metadiscoursal expressions

19
1. Learner Corpus Research keywords .
phraseology

• Function keywords in academic writing
• (Lee Chen 2009)
• 2 apprentice corpora
• Chinese undergraduate dissertations (Eng. Lang.)
• Comparable L1 student corpus from BAWE
• Both compared with expert corpus of journal
  articles from same field
• Key finding (article usage)
• e.g. the students in this study
• students will be more motivated if.

20
2. ESAP Corpus Research keywords . phraseology

• Content keywords in Problem-Solution text
  (Flowerdew 2008)
• 2 corpora of around 225,000 words each
• Learner corpus of recommendation-based reports
• Expert corpus of professional reports
• (analogue rather than exemplar corpusTribble
  2002)
• Key finding
• Overuse of superordinate terms in topic sentences
• e.g. find a solution to the problem
• Over-reliance on rubrics from assignment
  (lexical teddy bears, Hasselgren 1994)

21

• Cf. Paquots (2010) six interlanguage features
• Limited lexical repertoire
• Lack of register awareness
• Unidiomatic phraseology
• Semantic misuse (problem vs. issue or question)
• Overuse of connective devices
• Strings of connectives in subject-initial
  position

22
ESAP Learner Research metadiscourse

• Hylands (2005) model of metadiscourse
• Interactive
• guide reader through text using transitions
  etc.
• Interactional
• involves reader in text using hedges,
  boosters,
• attitude markers

23
1. ESAP Corpus Research Interactional

• Anticipatory it in student and published
  writing (Hewings Hewings 2002)
• 2 corpora of business writing
• Student corpus 15 MBA dissertations by NNS
  (123,633 words)
• Comparable corpus 28 papers from three different
  business studies journals (203,389 words )

24
1. ESAP Corpus Research Interactional

• Procedures
• Excluded it when propositional content, or
  text-organising role
• four categories derived from data
• Hedges, attitude markers, emphatics, attribution
• Results normalised per 1,000 words
• Results explanations
• Learners made less use of it clauses in hedging
• Greater use of it in other 3 categories
• Speculate more overt effort at persuasion on
  account of readership

25
2. ESAP Corpus Research Interactive

• Use of we in a learner corpus of reports (Luzon
  2009)
• 2 corpora
• Learner corpus of reports written by Spanish
  engineering students
• Corpus for comparison (analogue) engineering RAs
• Key finding
• Lack of awareness of conventions for we
• e.g. With this paper we want to give you some
  recommendations
• We are going to consider the advantages
  and

26
Current and future directions
27
Very recent initiatives in Learner Corpora

• Aggregate vs. individual data (cf. Hong article
  on software in Tono et al. 2012, ICCI project)
• Longitudinal studies (Meunier Littre 2013)
• More individual metadata
• MUCH longitudinal corpus (Eriksson et al.)

28
Very recent initiatives in Learner Corpora

• Domain- and genre-specific
• VESPA (Paquot) , CALE (Callies)
• Linguistic theories (genre, SFL, pragmatics,
  politeness strategies, cognitive linguistics etc.
• Computer-mediated, synchronous asynchronous
  (McDonald et al. 2013)

29

• Thank you