A Language Environment for Second Language Writers

1
A Language Environment for Second Language Writers

• Ola Knutsson
• KTH Nada

2
Outline

• Computer Assisted Language Learning
• Second Language Learning
• Language Technology in CALL
• A Field Study
• Grim a language environment
• Future directions and discussion

3
Why use computers in language learning?

• Old things with new tools?
• Warschauer Healey (1998)
• multimodal practice with feedback
• individualization in a large class
• variety in the resources available and learning
  styles used
• exploratory learning with large amounts of
  language data

4
Basic Components in Second Language Acquisition
(Gass 1997 Chapelle, 1998)

• Input (of target language)
• Apperception (noticing aspects in TL)
• Comprehension (semantic -syntax)
• Intake (semantic syntax)
• Integration (into learners interlanguage)
• Output (L2 production)

5
Second Language Learning Processes

• The acquisition and development of a second
  language is regarded as a complex processes
  requiring the interplay of motivation, identity,
  context, culture, intellectual competence
  (Sjögren, 1996)
• Second-language learning is viewed as a
  combination of spontaneous, inductive learning
  with systematic, deductive learning strategies
  (Laurillard, 1993)

6
Writing in the acquisition of second language

• Writing turns speech and language into objects of
  reflection and analysis (Vygotsky, 1962 Luria,
  1976)
• Far from transcribing speech, writing creates the
  categories in terms of which we become conscious
  of speech (Olson, 1995)

7
Possibilities and Risks with Language Technology

• Language technology can provide relevant
  feedback on the users unconstrained speech- or
  writing production.
• False feedback can fool the user
• Lack of feedback can mislead the user

8
Early 90s revisited (Severinson Eklundh et al,
1992)

• Language tools should support the writing
  process.
• Language processor instead of word processor
  (character processing).
• Implement and use language tools, in spite of
  their limitations.

9
Language tools wanted (Cedergren Severinson
Eklundh, 1992)

• Interactive concordances
• Interactive dictionaries
• Language based search function
• Swedish grammatical analysis
• A grammar checker for Swedish
• Language based editing

10
Language tools and efficient algorithms developed
at KTH Nada

• Stava (Kann Hollman, 1992)
• Granska (1998-?
• CrossCheck (2002-?) (incl. ProbGranska and
  AutoGranska)
• GTA a shallow parser for Swedish (2003)
• Other tools based on these programs and research
  insights

11
Point of departure

• Study the use of the grammar checker Granska
• Study the use of language technology
• Second Language Learners (students), in
  collaboration with teachers.

12
Why focus on errors?

• The learners errors are a register of their
  current perspective of the target language
  (James, 1998).
• Noticing a problem pushes the learner to
  modify his/her output (Swain Lapkin, 1995)
  syntactic processing mode important for the
  development of IL

13
Learning to write or learning a language?

• Learning to write in a second language?
• Learning a second language through writing (and
  reading)?

14
Granska

• Is a Swedish grammar checker developed at our
  department.
• Combines statistical and rule based methods.
• Contains a lot of opportunities for new
  applications.
• Different user interfaces

15
detection
diagnosis
correction
16
Language Tools for Learners

• How can we support second language learners
  writing with other technology than a robust
  grammar checker?
• What functionality is important in a learning
  environment for second language learners of
  Swedish?

17
User studies at a Swedish University

• Study 1
• Students Swedish as a foreign language, 7
  users, argumentative texts, 2 months
• Study 2
• Students Swedish as a foreign language, 11
  users. Different kinds of text genres, 4 months.

18
Use Granska whenever you want, and when you find
it appropriate

• Classroom observations
• Collecting students texts
• Collecting students judgments and comments on
  the feedback from Granska
• Questionnaires
• Interviews with students and teacher
• Final Workshop

19
Results

• Lack of feedback and misleading feedback
• Different sources of linguistic information
• - one tool is not enough
• Focus on form

20
More results

• Trusting the program
• Metalanguage and grammatical knowledge
• Transparency
• Interaction and integration

21
Some Insights

• One group of students wants to learn from the
  program.
• Another group of students wants to succeed in
  their writing.
• The teacher is very important as a mediator of
  the programs behavior and content.

22
How should we design an environment supporting
learning activities?

• Our proposal is a program called Grim.
• Grim is a environment with several language
  tools.
• Grim gives feedback on different aspects of the
  learners language.

23
Functionality in Grim

• Grim includes a simple word processor.
• Grim gives feedback on errors by using Granska.
• Grim provides grammatical information by using
  Granska Text Analyzer.
• Grim has an interface for concordances in 20
  million words corpus.
• Grim includes a dictionary with 8 language pairs.

24
Concluding remarks

• It is important to support different aspects of
  language learning.
• One robust grammar checker is not enough
• A ensemble of tools reduces the limitations of
  the individual tool

25
Future directions

• User studies with Grim
• Including communicational goals

26
Discussion

• Is Grim a learning environment?
• Should pedagogical packages be included in the
  environment?
• Focus on meaning?
• Why are computers underused in language teaching?
• How can we study the use of Grim in naturalistic
  settings without promoting it?

27
www.nada.kth.se/grim

• More than 1600 users