Schedule: 2009-09-10 (12:00 - 13:30)
Symposium: Frameworks for ICALL (Room A-32)

Title: Frameworks for ICALL
Authors: Cornelia Tschichold, Peter Wood, Mat Schulze
Abstract: The proposed symposium on frameworks for ICALL will take a wide-angled view on methods developed in computational linguistics and how these could be used to develop more intelligent, adaptable, and responsive CALL. Most of today’s CALL activities can be said to fall under the “computer as tool” category, while the SIG for NLP in CALL has always concentrated on the “computer as tutor” promise that NLP (Natural Language Processing) offers. This year’s symposium will focus on frameworks which can guide CALL researchers towards more advanced tools for autonomous learning.

The first paper takes a critical look at the constraints in force on any ICALL system:
While developments in the area of NLP over the last 15 years have certainly been considerable, John Nerbonne's statement that advances in NLP have surprisingly little impact on CALL, still holds true today. It can be argued that addressing the technological challenges for ICALL can provide only a partial answer to the question on why this is is the case. Perceptions of what learning a second language entails, how learning objects are to be administered and which kind of interventions are called for in a certain situation differ greatly not only among learners, instructors and developers, but also within these groups.
In this session the limits of independent learning within an ICALL framework will be discussed, concentrating not only on the technological limits, but also on limits that exist on other levels such as: institutions, pedagogies and learner beliefs.

The second paper on ICALL in Task-Based Language Teaching will consider the role of parser-based technologies in task-based language teaching (TBLT) (Ellis 2003, Willis 1996, Edwards & Willis 2005). Questions such as:
- What ICALL design and functionality is most suitable for pre-task and task-planning activities, during the work on the task itself, and for post-task activities?
- What task types lend themselves well to an ICALL implementation?
- What are the implications for the conceptualisation and design of ICALL systems?
will be addressed, looking for synergies of ICALL and TBLT. Most importantly, the design and employment of ICALL systems in a cognitively oriented TBLT framework will help ICALL to move away from its (perceived) link with more behavouristic views of language learning.

The third paper will focus on one of the core resources needed for ICALL, i.e. lexical databases.
Vocabulary could be considered to be the basis of language learning, and the orthographic word still functions as the central unit in NLP, yet the vocabulary trainers we have today are hardly the cutting-edge tools learners would love to use. A fundamental problem seems to be the underlying rather simplistic view of vocabulary knowledge.
To make better vocabulary trainers possible, we need lexical databases that contain not just more detailed information, but also more highly structured data, so that vocabulary practice can follow a didactically sensible progression and provide several examples of the word in varying contexts each at the right level for the learner. A template for such a dedicated lexical database will be sketched.

This symposium on NLP in CALL is organized by the EuroCALL Special Interest Group NLP in CALL.
Keywords: ICALL, NLP, computer as tutor, constraints, task-based language teaching, vocabulary learning

Main topic: Research in new language learning environments

Biodata: Mathias Schulze is an Associate Professor of German at the University of Waterloo, Canada, and the director of the Waterloo Centre for German Studies. He has published on ICALL and grammar acquisition and co-wrote the book 'Errors and Intelligence in CALL' (Heift and Schulze, 2007). Peter Wood is a PhD candidate at the university of Waterloo developing a n ICALL application to help learners of German acquire new vocabulary and word formation rules. He works at the Max Planck Institute for Psycholinguistics developing language archiving software Cornelia Tschichold is a lecturer in Applied Linguistics at Swansea University (Wales). She worked on a grammar checker for non-native speakers, and on the treatment of English multi-word units in computational lexicography. Her research interests include the acquisition of vocabulary and computer-assisted language learning (CALL).
Type of presentation Symposia

Paper category (Other)

Target educational sector Higher education

Language of delivery English

EU-funded project No

Schedule:	2009-09-10 (12:00 - 13:30) Symposium: Frameworks for ICALL (Room A-32)
Title:	Frameworks for ICALL
Authors:	Cornelia Tschichold, Peter Wood, Mat Schulze
Abstract:	The proposed symposium on frameworks for ICALL will take a wide-angled view on methods developed in computational linguistics and how these could be used to develop more intelligent, adaptable, and responsive CALL. Most of today’s CALL activities can be said to fall under the “computer as tool” category, while the SIG for NLP in CALL has always concentrated on the “computer as tutor” promise that NLP (Natural Language Processing) offers. This year’s symposium will focus on frameworks which can guide CALL researchers towards more advanced tools for autonomous learning. The first paper takes a critical look at the constraints in force on any ICALL system: While developments in the area of NLP over the last 15 years have certainly been considerable, John Nerbonne's statement that advances in NLP have surprisingly little impact on CALL, still holds true today. It can be argued that addressing the technological challenges for ICALL can provide only a partial answer to the question on why this is is the case. Perceptions of what learning a second language entails, how learning objects are to be administered and which kind of interventions are called for in a certain situation differ greatly not only among learners, instructors and developers, but also within these groups. In this session the limits of independent learning within an ICALL framework will be discussed, concentrating not only on the technological limits, but also on limits that exist on other levels such as: institutions, pedagogies and learner beliefs. The second paper on ICALL in Task-Based Language Teaching will consider the role of parser-based technologies in task-based language teaching (TBLT) (Ellis 2003, Willis 1996, Edwards & Willis 2005). Questions such as: - What ICALL design and functionality is most suitable for pre-task and task-planning activities, during the work on the task itself, and for post-task activities? - What task types lend themselves well to an ICALL implementation? - What are the implications for the conceptualisation and design of ICALL systems? will be addressed, looking for synergies of ICALL and TBLT. Most importantly, the design and employment of ICALL systems in a cognitively oriented TBLT framework will help ICALL to move away from its (perceived) link with more behavouristic views of language learning. The third paper will focus on one of the core resources needed for ICALL, i.e. lexical databases. Vocabulary could be considered to be the basis of language learning, and the orthographic word still functions as the central unit in NLP, yet the vocabulary trainers we have today are hardly the cutting-edge tools learners would love to use. A fundamental problem seems to be the underlying rather simplistic view of vocabulary knowledge. To make better vocabulary trainers possible, we need lexical databases that contain not just more detailed information, but also more highly structured data, so that vocabulary practice can follow a didactically sensible progression and provide several examples of the word in varying contexts each at the right level for the learner. A template for such a dedicated lexical database will be sketched. This symposium on NLP in CALL is organized by the EuroCALL Special Interest Group NLP in CALL.
Keywords:	ICALL, NLP, computer as tutor, constraints, task-based language teaching, vocabulary learning
Main topic:	Research in new language learning environments
Biodata:	Mathias Schulze is an Associate Professor of German at the University of Waterloo, Canada, and the director of the Waterloo Centre for German Studies. He has published on ICALL and grammar acquisition and co-wrote the book 'Errors and Intelligence in CALL' (Heift and Schulze, 2007). Peter Wood is a PhD candidate at the university of Waterloo developing a n ICALL application to help learners of German acquire new vocabulary and word formation rules. He works at the Max Planck Institute for Psycholinguistics developing language archiving software Cornelia Tschichold is a lecturer in Applied Linguistics at Swansea University (Wales). She worked on a grammar checker for non-native speakers, and on the treatment of English multi-word units in computational lexicography. Her research interests include the acquisition of vocabulary and computer-assisted language learning (CALL).
Type of presentation	Symposia
Paper category	(Other)
Target educational sector	Higher education
Language of delivery	English
EU-funded project	No