Spoken%20multimedia%20corpora%20for%20pedagogical%20purposes

1
Spoken multimedia corpora for pedagogical purposes
Birmingham Corpus Linguistics Conference 2007

• Sabine Braun (University of Surrey)
• Pascual Pérez-Paredes (Universidad de Murcia)
• Ylva Berglund (Oxford University)

2
Introduction

• The usefulness of corpora in language pedagogy is
  widely recognised.
• But there is a need for pedagogically relevant
  corpora, reflected e.g. in initiatives to create
  'ad-hoc' corpora in pedagogical contexts.
• The creation of pedagogically relevant corpora
  raises challenges for corpus design.
• Past and current initiatives have largely
  focussed on written corpora spoken discourse is
  becoming more important in pedagogical contexts.
• The creation of pedagogically relevant spoken
  corpora raises additional challenges for corpus
  design.

3
The challenges (1)

• Corpora contain textual records of discourse
  their interpretation requires (re-)contextualisati
  on.
• Learners may have difficulties analysing corpus
  data they require pedagogical mediation.
• Pedagogical corpus uses differ from linguistic
  description this requires e.g. pedagogically
  motivated query options.
• Corpora need to be integrated with curricula
  this requires e.g. complementarity of content and
  effective delivery.

• CORPUS DESIGNTraditional reference corpora
  (content, size, data format,transcription,
  annotation, query)

• CORPUS EXPLOITATIONData-Driven Learning (focus
  on non-linear reading concordances and co-texts)
  

Do not fully support pedagogical requirements.
4
The challenges (2)

• Spoken discourse is more dependent on shared
  physical contexts.
• It is adjusted to aural and online perception
  (e.g. chunking)
• It is affected by limitations of processing
  capacity (false starts, repair).
• It is marked by accents.
• It is multimodal.

• CORPUS DESIGNTraditionally representation in
  written format

• CORPUS EXPLOITATIONWork with text-only data and
  e.g. conversational markup

Again, this does not fully support pedagogical
requirements.
5
Requirements

• Format multimedia to retain multimodal character
  of spoken language
• Content complementary with curriculum topics,
  more coherence than in traditional corpora
• Pedagogically motivated transcription, annotation
  and alignment (transcript-video)
• Combination of query methods text-based
  exploration and application of corpus techniques
• Pedagogical enrichment of corpora with
  complementary resources (e.g. exercises,
  explanations)
• Effective delivery of corpora and additional
  resources to learners/teachers

6
Corpus creation (1)

• Examples ELISA and SACODEYL
• Interview format
• Video clips with transcript

• ELISA
• Professional English
• Accounts of professional life
• Different varieties

• SACODEYL
• 7 European languages
• Youth language corpora
• Speakers 13-15 and 16-18

• Communicatively relevant topics, e.g. in SACODEYL
  topics outlined in the Common European Framework
• Elicitation process briefing informants and
  prompting them during the interview, ensuring
  naturally flowing discourse

7
Corpus creation (1)
Example of topics in SACODEYL
Topic Interview questions Age CEF Gramm. functions
Holidays Where did you spend your last holidays? 13-15 16-18 A2 can describe past activities, personal experiences Past tense
What are your plans for the next holidays? 13-15 B1 can describe dreams, hopes and ambitions Future Conditonal Modal verbs
Plans for the future What are your plans for your career? 16-18 B1 can explain/give reasons for my plans, intentions and actions Future
On what grounds do you decide? 16-18 B2 can speculate about causes, consequences, hypothetical situations Conditional Modal verbs
8
Corpus creation (2)
CONTINUUM RAW, ORTHOGRAPHIC TRANSCRIPTION
ANNOTATED CORPORA
Transcription
TEI-compliant corpora
Markup
Pedagogic annotation
XML files
9
Corpus creation (2)
Transcription
SACODEYL TRANSCRIPTOR
TEI-compliant corpora
Markup
SACODEYL ANNOTATOR
Pedagogic annotation
XML files
10
Corpus creation (3)
SACODEYL TRANSCRIPTOR
11
Corpus creation (2)
LanguageES MediaFileNameES02.avi Participants
personChico name role Entrevistado
sex Hombre age 16 description person E
name Andrés Mercader Rodríguez role
Entrevistador sex Hombre age 32
description /METADATA

• METADATA
• Title La Unión Europea une a los ciudadanos
• Date Recording2006-11-05
• Date Transcription2007-02-02
• LocaleI.E.S. Floridablanca,Murcia, España
• Principal Investigator Pascual Perez-Paredes
• ResearcherPascual Perez-Paredes
• Transcriber Encarnación Tornero Valero
• Editor
• Autority SACODEYL Project
• ID

12
(No Transcript)
13
(No Transcript)
14
(No Transcript)
15
(No Transcript)
16
(No Transcript)
17
(No Transcript)
18
Corpus query

• Query options will support text- and corpus-based
  exploration and include e.g.
• Easy access to entire interviews
• A topic index supporting the analysis of similar
  sections across interviews ("topic concordances")
• Other indices based on the annotation categories
• Ready-made data (e.g. frequency lists of each
  interview selective concordances)
• A concordancer for extended/advanced search
  adapted to pedagogical requirements

19
Corpus query
20
Pedagogical enrichment

• The corpora will be enriched with prototypical
  learning activities.
• These will focus on one interview section or one
  interview as a whole or sections across
  interviews
• They will include e.g.
• linguistic and cultural explanations and
  exercises(form-focussed as well as
  communication-oriented),
• (listening) comprehension and production tasks,
• explorative tasks (concordance-based as well as
  interview-based).
• Use of authoring tool Telos Language Partner to
  create learning packages with ranges of
  activities.

21
Pedagogical enrichment
22
Pedagogical enrichment
23
Pedagogical enrichment
24
Pedagogical enrichment
25
Corpus delivery

• Effective delivery as a further prerequisite for
  integration into curriculum
• In SACODEYL, use of Moodle learning platform,
  giving access to
• Corpora (query interfaces)
• Resources created in the project (different types
  of learning activities)
• Resources created by future corpus users

26
Summary

• Method outlined is transferable to other
  pedagogical contexts, topics, languages
• Method helps to use corpora more efficiently in
  pedagogical contexts from sporadically used
  resource to systematic exploitation
• Corpus creation complies with standards to
  facilitate reuse of corpora for other contexts
  (research)
• ?

27
Contact

• Sabine Braun s.braun_at_surrey.ac.uk
• Pascual Pérez-Paredes sacodeyl_at_um.es
• Ylva Berglund ylva.berglund_at_oucs.ox.ac.uk
• And visit our poster session
• As well as our websites
• www.um.es/sacodeyl
• www.corpora4learning.net/elisa