Optimizing the role

of language in
Technology-Enhanced
Learning

Sylviane Granger (ed.)

Louvain-la-Neuve, 4-5 October 2007

 SCIENTIFIC COMMITTEE

Georges Antoniadis (Université Stendhal Grenoble, France)

Cédrick Fairon (Université catholique de Louvain, Belgium)
Sylviane Granger (Université catholique de Louvain, Belgium)
Adam Kilgarriff (Lexical Computing Ltd, United Kingdom)
Fanny Meunier (Université catholique de Louvain, Belgium)
John Nerbonne (Universiteit Groningen, The Netherlands)
Magali Paquot (Université catholique de Louvain, Belgium)
Claude Ponton (Université Stendhal Grenoble, France)
Michael Rundell (Lexicography Masterclass Ltd & Macmillan Dictionaries, United Kingdom)
Serge Verlinde (Katholieke Universiteit Leuven, Belgium)

ORGANIZING COMMITTEE

Centre for English Corpus Linguistics, Université catholique de Louvain (Belgium)

Sylviane Granger
Fanny Meunier
Sylvie De Cock
Valérie Decuyper
Céline Gouverneur
Claire Hugon
Marie-Aude Lefer
Magali Paquot
Jennifer Thewissen

Centre de traitement automatique du langage, Université catholique de Louvain (Belgium)

Cédrick Fairon
Bernadette Dehottay
Thomas François
 Table of Contents

Charles Alderson, Computer-adaptive language testing ........................................ 1

Eckhard Bick, VISL: A cross-language approach to NLP- and games-based
grammar teaching .............................................................................................. 5
Jozef Colpaert, CALL software design principles and the integration of NLP ...... 7
Piet Desmet and Hans Paulussen, CorpusCALL: Challenges and opportunities ... 9
Sylviane Granger, The contribution of learner corpus research to TELL ............. 13
Thomas Hansen, Feedback methods in computer assisted pronunciation
training applications using automatic speech recognition ............................... 17
Holger Hvelplund, Language technology projets at IDM ....................................... 19
Adam Kilgarriff, Using corpora in language learning: the Sketch Engine ........... 21
Thomas Koller, A plurilingual ICALL system for Romance languages ................. 25
Agnes Kukulka-Hulme, Like stars in the firmament: Language learning on
mobile devices ................................................................................................... 29
Marie-Noëlle Lamy, Computer-mediated communication for language
learning ............................................................................................................. 33
Claudia Leacock, Writing English as a second language: A proofreading tool .... 37
Fanny Meunier, Second language acquisition theory and TELL ........................... 41
John Nerbonne, Detecting syntactic interference ................................................... 43
Richard Pemberton, Planning a smart phone system to support self-directed L2
vocabulary learning ......................................................................................... 45
Michael Rundell, The dictionary of the future ......................................................... 49
Emma Shercliff, Macmillan English Campus: A case study .................................. 53
Serge Verlinde, The Base lexicale du français (BLF), a free Web-based
learning environment for French vocabulary ................................................... 57
Carl Vogel, Linguistic anomaly .............................................................................. 61
David Wible, Evolving approaches to Web-supported language learning: From
platforms to platform-independent tools ........................................................... 65
Johannes Widmann, The SACODEYL project – Corpus exploitation for
language learning purposes .............................................................................. 69
 Computer-adaptive language testing

Charles Alderson
Lancaster University

1. Introduction
Computer-adaptive testing (CAT) involves presenting learners with items thought to
be most suitable for them, and adjusting the selection of items in light of the learners’
responses to previous items. The classic case of CAT involves the construction of a
bank (collection) of test items which have been calibrated in terms of their empirical
difficulty. Learners are typically initially presented with an item of medium difficulty.
If their response is correct, they will then be presented with a more difficult item. If
their response is incorrect, they are then presented with an easier item. If their response
to the second item is correct they are given a more difficult item and if incorrect, an
easier item. The computer calculates the learner’s ability level (or score) on the fly as
well as the reliability of the test as administered up to that point. Items from the bank
are presented to test-takers following specially developed algorithms for the selection
of the initial test item, subsequent test items, and a rule for concluding the test – ie the
criteria to be met for the test to be terminated. Typically, the test is terminated when a
given level of reliability has been reached, or when a pre-determined number of items
has been delivered.
The advantages are that tests can be tailored to a learner’s ability level rather than
wasting time and effort by presenting them with items that are far too easy or far too
difficult. As a consequence, tests can be markedly shorter than traditional linear tests,
and thus more efficient. In addition, since each learner takes a different test than his or
her fellow test-takers, cheating is made much more difficult. The major disadvantages
are that in order to be able to predict a learner’s ability level items need to be pre-
tested and analysed using an Item Response Theory model – IRT (which allows the
estimation of a learner’s ability level independent of the difficulty of the items) – but
IRT requires relatively large numbers of pilot test candidates for reliable ability
estimates. Secondly, in high-stakes testing situations (like the TOEFL) learners are
often schooled in remembering which items they have taken, and the item bank can be
reconstructed if sufficient numbers of candidates recall the items (this has happened in
China, for example, where CAT versions of TOEFL were compromised). In addition,
truth-in-testing laws in some states of the USA mean that the test items constituting
the basis of the test-taker’s score have to be made available to test-takers on request.

Optimizing the role of language in Technology-Enhanced Learning (2007): 1-3

2 CH. ALDERSON

This inevitably compromises the test bank. As a result, ETS (the developers of
TOEFL) have ceased developing CATS since, as they put it, “feeding the CAT is too
expensive”.

2. Outline of the presentation/demo

In this talk I will briefly present the work I have done on CATs to date and then
discuss possible amendments to the design of CATs to make them more relevant to
learners and to test purposes. My main involvement with CATs, apart from critiquing
their value, has been on the DIALANG Project. DIALANG is a suite of Internet-
delivered diagnostic tests of 14 European languages, funded by the European Union,
and based on the Common European Framework. It contains tests of reading, listening,
writing, grammar and vocabulary, as well as a test of vocabulary size and a self
assessment battery. The current version of the test is adaptive at what we call the test
level, and I shall describe how this works and how feedback is given. Algorithms have
been developed – but not yet implemented – to make the test adaptive at the item level,
as in the classic case described above, and I shall describe the problems and solutions
involved in making a test, which is delivered over the Internet, adaptive at item level. I
shall then describe and discuss further developments of CATs. These are much less
likely to take place in the context of high-stakes proficiency tests, for the reasons given
above, but in the context of language learning, be that of progress or achievement tests
or, in the case I am interested in, in further exploration and refinement of the diagnosis
of learners’ strengths and weaknesses.
Adaptation to the learner’s response need not be simply on the basis of the difficulty of
the item responded to, but on the basis of item content. Thus, if the items in the CAT
bank are characterised not merely in terms of their empirical difficulty, but in terms of
the language features they test, or in terms of the skill or sub-skill they measure, then
one can envisage an adjustment of the CAT algorithm to take account of what is being
tested. Thus, a diagnostic CAT of one’s command of structures in the language could
select items on, for example, the use of the present perfect, and explore how
thoroughly a learner mastered that tense/ aspect in a variety of contexts, or with a
range of different verbs. Success on a range of items could lead to the selection of
items on a different aspect of syntax, or to the presentation of items in the same
syntactic area but known to be more “advanced” in terms of the acquisition sequence
and/or the development of the learner’s syntactic competence. Similarly, one could
envisage tests of vocabulary being structured according to the frequency of words in
the language, or according to particular semantic fields, or domains of use or register,
and so on. Computer adaptivity would then enable a more or less thorough exploration
of strengths and weaknesses in lexical knowledge.
Similarly, in tests designed to establish a learner’s level on the Common European
Framework of Reference or some similar relevant standard, CATS could present items
calibrated and standard-set at the different CEFR levels, and the degree of a learner’s
 COMPUTER-ADAPTIVE LANGUAGE TESTING 3

mastery of items in a given skill or language use domain at each particular level could
be explored in some depth.
Another adaptation of the principle of adaptivity could take account of learner
characteristics (age, mother tongue, years of learning, gender, topics of interest, area of
academic study, etc). The learner would select from a menu of possible characteristics
those that applied to them, and the computer would only present items known to be
suitable for learners with such a profile – or, indeed, items known to be a challenge for
such learners.
Finally, instead of the computer making the decision on which next item to select, the
learner could be allowed to do so (by, for example, requesting a more difficult item, or
one on a different linguistic feature or another topic or academic discipline).

3. Issues and challenges

The major challenge in the field is to identify relevant characteristics of items and of
learners which would provide meaningful diagnoses, or results relevant to further
learning, and this requires a much better theory of diagnosis and language
development than we currently possess. I hope in the discussion that we can explore
whether NLP techniques can contribute to this notion of adaptivity.

4. References
CHALHOUB-DEVILLE M. (ed) (2000), Computer-adaptive tests of reading, Cambridge, CUP.
CHALHOUB-DEVILLE M. and DEVILLE C. (1999), “Computer Adaptive Testing in Second
Language Contexts” in Annual Review of Applied Linguistics, 19: 273-299.
DUNKEL P.A. (1999), Considerations in Developing and Using Computer-Adaptive Tests to
Assess Second Language Proficiency, http://www.cal.org/resources/Digest/cat.html (last
accessed 7.8.07).
WAINER H. (ed.) (2000), Computer-Adaptive Testing: A Primer, Mahwah, NJ, USA,
Lawrence Erlbaum Associates.
 VISL: A cross-language approach to NLP-
and games-based grammar teaching

Eckhard Bick
University of Southern Denmark

1. Introduction
VISL (Visual Interactive Syntax Learning) is an integrated interactive user interface
for teaching grammatical analysis on the Internet, developed at the University of
Southern Denmark, offering a unified system of analysis for 25 different languages, 8
of which are supported by live grammatical analysis of running text. For reasons of
robustness, efficiency and correctness, the system’s internal tools are based on the
Constraint Grammar formalism (Karlsson 1990), but users are free to choose from a
variety of notational filters, supporting different descriptional paradigms, with a
current teaching focus on syntactic tree structures, language independent grammatical
categories and the form-function dichotomy. VISL’s core NLP-programs use the
author’s hybrid multi-level parsers (http://beta.visl.sdu.dk), while teaching applications
(http://visl.sdu.dk) and corpus searching tools (http://corp.hum.sdu.dk) are
implemented as platform independent Java-programs and Perl-cgi’s. Though lexica
and parsing rules are developed individually for each language, a common CG and
treebank data format facilitates source data transfer into grammar teaching games,
structural or color based visualisation, and linguistic revision of corpus data.

2. Outline of the presentation/demo

In a modern school or university environment, grammar teaching is often plagued by
the fact that the subject is perceived as “academic” and “uninteresting”, and its
inherent analytical view on language conflicts with a current language teaching focus
on assimilation, naturalness, communicative media etc. Also, grammar teaching is
affected by a cross-language handicap, because students are confronted with different
formal systems and terminology, depending on the individual language taught (e.g.
latinid vs. native terminology, morphological vs. functional word classes, syntax trees
for English, dependency grammar for Czech, topological fields for Danish). In order to
address these problems, the VISL system has introduced a novel, unified approach

Optimizing the role of language in Technology-Enhanced Learning (2007): 5-6

6 E. BICK

across languages, built on a clear distinction between function and form, and tied to
visually stable clues, such as iconic abbreviations, symbols and colour coding.
The presentation/demo will demonstrate how these principles can be implemented in
the form of internet-based grammar games such as WordFall, Labyrinth, Syntris etc.,
as well as tree structures and corpus tools.

3. Issues and challenges

While games and treebanks can be based on manually annotated data, a truly flexible
system and, not least, language teaching based on empirical, corpus-derived evidence
cannot realise its full potential without robust, automatic NLP, and even apparently
“closed” exercises and games become dependent on such tools if a higher degree of
lexical or structural variation is to be achieved, or where a teacher would like to adapt
exercises or games to a given text book tradition or recently treated literature. For
these reasons (and also for the sake of linguistically more robust interfaces), I believe
the integration of main stream parsing technology to be one of the major challenges in
the future development of CALL applications.

4. References
BICK E. (2005-1), “Grammar for Fun: IT-based Grammar Learning with VISL”, in
P.J. Henriksen (ed.), CALL for the Nordic Languages, København, Samfundslitteratur:
49-64 (Copenhagen Studies in Language).
BICK E. (2005-2), “Live use of Corpus data and Corpus annotation tools in CALL: Some new
developments in VISL”, in H. Holmboe (red.), Nordic Language Technology, Årbog for
Nordisk Sprogteknologisk Forskningsprogram 2000-2004 (Yearbook 2004), Copenhaguen,
Museum Tusculanum: 171-186.
DAVIES G. (ed.) (2007), Information and Communications Technology for Language
Teachers (ICT4LT), Slough, Thames Valley University (online: http://www.ict4lt.org/).
EUROCALL bibliography: http://www.eurocall-languages.org/resources/bibliography/books.
html
FITZPATRICK A. and DAVIES G. (eds) (2003), The Impact of Information and Communications
Technologies on the Teaching of Foreign Languages and on the Role of Teachers of
Foreign Languages.
KARLSSON et al. (1995), Constraint Grammar – A Language-Independent System for Parsing
Unrestricted Text, Mouton de Gruyter.
WARSCHAUER M. and HEALEY D. (1998), “Computers and language learning: An overview”,
in Language Teaching, 31: 57-71.
WARSCHAUER M. (1996), “Computer-assisted language learning: an introduction”, in
S. FOTOS (ed), Multimedia Language Teaching, Tokyo, Logos International.
 CALL software design principles
and the integration of NLP

Jozef Colpaert
University of Antwerp

1. Introduction
The role and shape of solutions for language learning should not be based on a
technology-driven (not even NLP-driven) approach, but on an accurate specification of
what is needed for a particular language learning situation. We therefore have to create
a language learning environment first, defined as an architecture of actors and
components and their mutual interactions, before deciding on the language method,
media, systems, technologies and NLP routines needed.
Our current research focuses on the implementation of Distributed Language Learning,
a conceptual and methodological framework (DLL) for designing language learning
solutions in distributed environments.

2. Outline of the presentation/demo

In this presentation we will explain the concept of DLL, and show how it has been
applied to system development, content structuring, course design and even to the
design of a completely new Language Institute.
In the case of system development, we will present a DLL-based software architecture
that allows the integration of NLP-routines (Heift and Schulze 2007) on the level of
error analysis and answer evaluation but also on the level of the interface (especially
for physically, visually or auditory challenged learners). We will show examples in 3D
game scripting, language testing and mobile learning.

3. Issues and challenges

Our current challenge is the design of an architecture for an intelligent server-based
tutoring system for mobile devices, which will be the topic of our next FP7 proposal.

Optimizing the role of language in Technology-Enhanced Learning (2007): 7-8

8 J. COLPAERT

4. References
COLPAERT J. (2007), “Distributed Language Learning”, editorial in Computer Assisted
Language Learning, Vol. 20, No. 1, February 2007: 1-3.
COLPAERT J. (2007). “Pedagogy-driven design for online language teaching and learning”, in
CALICO Journal 23:3: 477-497.
COLPAERT J. (2007). “Toward an ontological approach in goal-oriented language courseware
design and its implications for technology-independent content structuring”, in Computer
Assisted Language Learning, vol. 19, 2&3: 109-127.
COLPAERT J. (2004). Design of Online Interactive Language Courseware: Conceptualization,
Specification and Prototyping. Research into the impact of linguistic-didactic functionality
on software architecture. (Doctoral dissertation). University of Antwerp, 2004, 342 p. UMI
micropublication number 3141560. Also available on www.didascalia.be/doc-design.pdf.
HEIFT T. and SCHULZE M. (2007), Errors and Intelligence in Computer-Assisted Language
Learning. Parsers and Pedagogues, Milton Park (Routledge Studies in Computer-Assisted
Language Learning (ed. C. Chapelle).
 CorpusCALL: Challenges and opportunities

Piet Desmet and Hans Paulussen

K.U.Leuven Campus Kortrijk

1. Introduction
This talk is situated in the field of corpusCALL, the use of corpora within CALL
(Computer Assisted Language Learning), that has gained growing importance within
the CALL research community as can be seen from recent publications (Sinclair 2004,
Gavioli 2005, Braun et al. 2006, Chambers 2007), and the introduction of SIG
communities based on this theme within EuroCall & Calico.
Our research group is quite active within the field of corpusCALL: we have two
projects on this domain running at the moment. This should be placed in our general
interest in CALL, which has recently led to the foundation of ALT, Research Center on
CALL. Our current projects involve research topics such as harnessing collective
intelligence in e-learning environments, effectiveness of electronic learning platforms,
authoring systems for the creation of half-open and open supported tasks and
electronic language testing.

2. Outline of the presentation/demo

This talk consists of two parts. First of all, we will give an overview of the different
approaches of using corpora for foreign language learning, and CALL in particular.
The second part will deal with some aspects of corpus creation and the need of
standardisation. Both parts will use examples from different projects, including the
parallel corpus project REBECA (a collaborative project between K.U.Leuven
Campus Kortrijk and FUNDP, Namur) and the recently started DPC project. The
Dutch Parallel Corpus project (DPC) is a STEVIN project, organised by a consortium
of Dutch and Flemish universities and translation institutes, which aims at compiling a
multilingual multifunctional corpus for language technology, translation studies,
linguistics and corpusCALL.

Corpora have been created and explored for a long time for different purposes
including language technology and linguistics. Only the last ten years has corpus
exploration moved to other domains, including foreign language learning and CALL
(Computer Assisted Language Learning). Moving away from language specialists to

Optimizing the role of language in Technology-Enhanced Learning (2007): 9-11

10 P. DESMET AND H. PAULUSSEN

general language users, corpus exploitation requires an adapted approach which

demands different exploitation tools and high quality annotation. We will show where
corpora can be useful in language teaching and explain the quality requirements for
corpusCALL.
The creation of a corpus has improved considerably over the last ten years, due to the
ever growing computer capacity, the interconnectivity between computers of different
platform types and the introduction of the internet to the general public. Moreover,
most texts are nowadays created electronically. These technological improvements
show that corpus creation has become a very easy task, at least as far as collecting text
samples is concerned. However, cleaning, structuring and annotating corpora requires
careful attention, especially when qualitative exploitation is the ultimate goal. An
important improvement in corpus compilation and further exploitation is the
introduction of character standardisation (Unicode) and document standardisation via
XML (e.g. TEI and XCES). Although the XML formats require specific handling,
their importance in compiling and distributing text corpora cannot be underestimated.
The advantages of XML distribution will be illustrated.

3. Issues and challenges

In the context of pure NLP applications (e.g. machine translation), corpora are mainly
used as linguistic resources to feed a particular application. The corpus is usually
transformed into meaningful chunks complying with the requirements of some
statistical application. The corpus itself remains invisible to the outside world. In the
context of language learning, on the other hand, corpora remain very “visible” to the
end-user. Therefore, higher quality standards are required for corpus compilation,
annotation and exploitation.
The main challenges in the use of corpora for language learning are situated in further
exploitation of corpora. In foreign language learning, corpora can be explored in at
least three different stages with reference to the language learning process: (i) corpus
extracts in the preparation of language material; (ii) corpus samples during the learning
activity and (iii) corpus samples used as feedback after the learning activity. The use
of standardised XML formats can improve the exploitation of corpora in each stage.

4. References
BRAUN S., KOHN K. and MUKHERJEE J. (2006), “Corpus technology and language pedagogy”,
in English Corpus Linguistics, Vol. 3, Frankfurt am Rain, Peter Lang.
CHAMBERS A. (2007), Integrating Corpora in Language Learning and Teaching. Special
Issue of ReCALL, Volume 17(3).
DESMET P. and HÉROGUEL A. (2005), “Les enjeux de la création d’un environnement
d’apprentissage électronique axé sur la compréhension orale à l’aide du système auteur
IDIOMA-TIC”, in ALSIC (Apprentissage des Langues et Systèmes d’Information et de
Communication), 8. http://alsic.u-strasbg.fr/v08/desmet/alsic_v08_12-poi4.htm
 CORPUSCALL: CHALLENGES AND OPPORTUNITIES 11

DESMET P. (2006), “L’apprentissage/enseignement des langues à l’ère du numérique:

tendances récentes et défis”, in Revue française de linguistique appliquée, 11: 119-138.
DESMET P. and EGGERMONT C. (2006), “FRANEL: Un environnement électronique
d’apprentissage du français qui intègre des matériaux audio-visuels et qui est à la portée de
tous”, in Cahiers F. Revue de didactique français langue étrangère , 7: 39-54.
DESMET P. (2007), “L’apport des TIC à la mise en place d’un dispositif d’apprentissage des
langues centré sur l’apprenant”, In I.T.L. International Journal of Applied Linguistics 153
(in press).
DEVILLE G., DUMORTIER L. and PAULUSSEN H. (2004), “Génération de corpus multilingues
dans la mise en oeuvre d’un outil en ligne d’aide à la lecture de textes en langue
étrangère”, in G. Purnelle, C. Fairon and A. Dister (eds.), Le poids des mots, Actes des 7es
journées internationales d’analyse statistique des données textuelles, JADT 2004,
Louvain-la-Neuve, March 2004: 304-312.
GAVIOLI L. (2005), Exploring corpora for ESP learning, Amsterdam, John Benjamins.
MACKEN L., TRUSHKINA J., PAULUSSEN H., RURA L., DESMET P. and VANDEWEGHE W.
(2007), “Dutch Parallel Corpus: a multilingual annotated corpus”, in Proceedings of The
fourth Corpus Linguistics conference, University of Birmingham.
SINCLAIR J. McH. (2004), How to use corpora in language learning, Amsterdam, John
Benjamins.

Website ALT, Research Center on CALL: www.kuleuven-kortrijk.be/ALT

Website LINGUATIC project: www.kuleuven-kortrijk.be/linguatic
Website DPC project: www.kuleuven-kortrijk.be/dpc
 The contribution of learner corpus research to
TELL

Sylviane Granger
Université catholique de Louvain

1. Introduction
Learner corpus research is a fairly young but highly dynamic research field that
emerged in the late 1980s. It focuses on the collection, annotation and computer-aided
analysis of vast electronic collections of authentic written and spoken data produced
by foreign language learners. The Centre for English Corpus Linguistics of the
University of Louvain (UCL) has played a key role in shaping the field and
demonstrating its tremendous pedagogical potential. In my presentation I will briefly
describe the work carried out at Louvain and sketch the numerous possibilities it offers
for Technology-Enhanced Language Learning.

2. Outline of the presentation/demo

LEARNER CORPUS COLLECTION
The learner corpora collected at Louvain contain data produced by foreign language
learners of English and French. One of the characteristics of the corpora that
distinguishes them from other similar collections is that they contain data from a wide
range of learner populations. The International Corpus of Learner English (ICLE) is a
corpus of argumentative essays produced by higher intermediate to advanced learners
from 16 different mother tongue backgrounds. The Louvain International Database of
Spoken English Interlanguage (LINDSEI) is the spoken counterpart of the ICLE and
currently covers 11 mother tongue backgrounds. The French Interlanguage Database
(FRIDA) contains written data from two well-defined learner populations: English-
and Dutch-speaking learners of French plus a mixed subcorpus representing a wide
range of mother tongue backgrounds. One important characteristic of our corpora is
that they are richly documented. In ICLE and LINDSEI over 20 task and learner
variables have been recorded for each of the texts through a detailed profile
questionnaire that all learners were requested to complete. All the variables have been
stored in a database and can be used by researchers as queries to compile subcorpora
that match certain criteria, thus allowing for interesting comparisons (German- vs.
Spanish-speaking learners, etc.).

Optimizing the role of language in Technology-Enhanced Learning (2007): 13-15

14 S. GRANGER

LEARNER CORPUS ANALYSIS

Two methods of analysis – Contrastive Interlanguage Analysis (CIA) and Computer-
Aided Error Analysis (CEA) – have been very popular among learner corpus
researchers. CIA is a very powerful, fully automatic heuristic that uncovers the
patterns of overuse, underuse and misuse that distinguish learner writing or speech
from native or expert user data. CEA is a highly time-consuming but extremely fruitful
process that enables researchers to have access to comprehensive catalogues of errors
for a given learner population. In my presentation I will illustrate the two methods and
describe the systems of error annotation we have designed to make errors amenable to
subsequent automated processing.
PEDAGOGICAL APPLICATIONS
Our learner corpus work has informed a range of pedagogical applications:
- Exercises in the FreeText CALL program for learners of French as a Foreign
Language (Granger 2003);
- ‘Get it right’ notes and ‘Improve your writing skills’ section in the new edition of
the Macmillan English Dictionary for Advanced Learners (Rundell and Granger
2007; Gilquin et al. in press)
- Web-based error interface: Exxelant (Granger et al. 2007).

3. Issues and challenges

Although learner corpus research has already generated some useful pedagogical
applications, its potential is much greater and the extent of learner corpus integration
into future applications is likely to increase in the near future. The following are but
some of the many avenues for future learner-corpus-informed research: computer-
adaptive testing, fleshing out of the descriptors of the Common European Framework
levels, error detection and modeling (track what learners have mastered and what they
still need to learn; provide appropriate feedback), TELL materials design (notably
electronic dictionaries and grammars), incorporation of learner corpus collection and
annotation into TELL mobile/web-based environments, speech recognition. All these
applications call for synergies between specialists in a wide range of disciplines: TEL,
language teaching, language testing, NLP, corpus linguistics and second language
acquisition.

4. References
BELZ J.A. and VYATKINA, N. (2005), “Learner Corpus Research and the Development of L2
Pragmatic Competence in Networked Intercultural Language Study: The Case of German
Modal Particles”, in Canadian Modern Language Review 62.1: 17-48.
GRANGER S. (ed.) (1998), Learner English on Computer, Addison Wesley Longman, London
and New York.
 THE CONTRIBUTION OF LEARNER CORPUS RESEARCH TO TELL 15

GRANGER S. (2003), “Error-tagged learner corpora and CALL: a promising synergy”, in

CALICO (special issue on Error Analysis and Error Correction in Computer-Assisted
Language Learning) 20(3): 465-480.
GRANGER S., DAGNEAUX E. and MEUNIER F. (2002), The International Corpus of Learner
English. CD-ROM and Handbook, Presses universitaires de Louvain, Louvain-la-Neuve.
Available from http://www.i6doc.com
GILQUIN G., GRANGER S. and PAQUOT M. (in press), “Learner corpora: the missing link in
EAP pedagogy”, in Journal of English for Academic Purposes 6 (4).
GRANGER S., HUNG J. & PETCH-TYSON S. (eds.) (2002), Computer Learner Corpora, Second
Language Acquisition and Foreign Language Teaching, Benjamins, Amsterdam and
Philadelphia.
GRANGER S., KRAIF O., PONTON C., ANTONIADIS G. and ZAMPA V. (2007), “Integrating
learner corpora and natural language processing: A crucial step towards reconciling
technological sophistication and pedagogical effectiveness”, in ReCALL 19(3): 252-268.
RUNDELL M. and GRANGER S. (2007), “From corpora to confidence”, in English Teaching
Professional, 50: 15-18. Reprinted in MED Magazine, 46, August 2007 http://www.
macmillandictionary .com/ MED-Magazine/August2007/46-Feature_CorporatoC.htm.
NESSELHAUF N. (2004), “Learner corpora and their potential for language teaching”, in J.
Sinclair (ed.), How to use corpora in language teaching, Benjamins, Amsterdam and
Philadelphia: 125-152.
WIBLE D., KUO C.-H., CHIEN F.-Y., LIU A. and TSAO N.-L. (2001). “A web-based EFL
writing environment: integrating information for learners, teachers, and researchers”, in
Computers and Education, 37: 297-315.
 Feedback methods in computer assisted
pronunciation training applications using automatic
speech recognition

Thomas Hansen
University of Southern Denmark

1. Introduction
The field of Computer Assisted Pronunciation Training (CAPT) has seen an explosion
in the use of Automatic Speech Recognition (ASR) technology within the past two
decades. Contemporary applications come equipped with commercial battle cries of
success that leaves one wondering why the use of such applications is not more
widespread than it is and also why second language acquisition (SLA) so often still
fails.
The main question in this connection is whether the feedback strategies which are
presently employed work or how they should be structured to maximize learner
benefit.

2. Outline of presentation/demo
Contemporary CAPT applications employ a variety of feedback methods, of which the
pedagogical value will be discussed. A potential strategy, or roadmap, for improving
the effectiveness of ASR in CAPT applications will be outlined for discussion.

3. Issues and challenges

Shaping and developing Automatic Speech Recognition technology in such a fashion
that a more detailed and constructive level of feedback can be achieved.

4. References
HANSEN Th. (2006), “The Four K’s of feedback?” In Proceedings of the 4th International
Conference on Multimedia and Information and Communication Technologies in
Education (m-ICTE2006), Seville: 342-346.
BERNSEN N.O., HANSEN Th., KIILERICH S., MADSEN T. (2006), “Field Evaluation of a Single-
Word Pronunciation Training System”, in Proceedings of The Fifth International

Optimizing the role of language in Technology-Enhanced Learning (2007): 17-18

 TH. HANSEN 18

Conference on Language Resources and Evaluation, LREC 2006, Genova, May, 2006:
2068-2073.
HINCKS R. (2005), Computer Support for Learners of Spoken English, Doctoral dissertation,
School of Computer Science and Communication.
HINCKS, R. (2003), “Speech technologies for pronunciation feedback and evaluation”, in
ReCall: 3-20.
NERI A., CUCCHIARINI C. and STRIK H. (2006), “ASR corrective feedback on pronunciation:
Does it really work?”, in Proceedings of ICSLP2006, Pittsburgh: 1982-1985.
 Language technology projects at IDM

Holger Hvelplund
IDM, Paris

1. Introduction
Based on experience with production of electronic versions of monolingual and
bilingual dictionaries the emphasis in the presentation will be on how content for cross
media products can be produced efficiently, how and why different types of content
can/should be integrated; how content can be accessed and adapted in different ways
depending on the user, the medium, and the context the content is used in.

2. Outline of the presentation/demo

Short demonstration of:
• DPS – a tool for compiling content for different medias and target audiences.
• A few ELT dictionary products with examples of:
• Different ways of integration language teaching content with the content of the
dictionary.
• How the dictionary can produce content that can be used by language teaching
components.
• Publishing same content on different medias and for different user audiences.
• Teacher resource database.

3. Issues and challenges

• Cost efficient production of content for different medias and target audiences.
• Integration of services in new ways where potential in new technologies are
exploited. For example, providing language teaching services with Skype, podcast,
broadband (like in services from Praxis Language).
• Intelligent user interface with personalized and integrated features.

Optimizing the role of language in Technology-Enhanced Learning (2007): 19-20

 H. HVELPLUND 20

4. References
Recent productions from IDM:
• DPS – Dictionary Production System.
• Production of dictionaries with XDCC for publishers like
- Oxford University Press (including CD-ROM version of Oxford Advanced Learner’s
Dictionary).
- Pearson Education (including CD-ROM and online version of Longman Dictionary of
Contemporary English).
- Macmillan Dictionaries (including CD-ROM version of Macmillan English Dictionary).
- Cambridge University Press (including CD-ROM and online version of Cambridge
Advanced Learner’s Dictinary and CD-ROM version of Cambridge Grammar of
English).
- + several others.
 Using corpora in language learning:
the Sketch Engine

Adam Kilgarriff
Lexical Computing Ltd

1. Introduction
I am writing an essay about my career plans, and I want to talk about goals. How does
the word work? What sorts of sentences might I construct around it, with what
collocates?
The current range of EFL dictionaries aim to help, and are well-designed, sophisticated
tools which specify grammatical patterns and collocates, and show the user a range of
example sentences. Often that will be enough. But they are limited to a couple of
column inches for a word like goal (in which they must cover all of its meanings) and
sometimes they just do not cover the case the student is interested in. When that
happens, where should they go next?
It is tempting to say that they should go and look in the corpus: after all, that is where
the people who wrote the dictionary went. The idea has been discussed at length in the
“Teaching and Language Corpora” community. The problem is that reading
concordance lines is a skill requiring advanced language competence, and is simply
offputting to most learners. The issue may be presented as follows: the dictionary is a
highly condensed short summary of the word’s behaviour. The corpus is the raw data
for such a summary, not at all condensed or summarized. The user would like a point
in between: not as short and minimal as the dictionary, but with a level of abstraction
and generalization.
Using techniques from computational linguistics, we have applied just this logic to
produce ‘word sketches’ – one-page accounts of the grammatical and collocational
behavior of word, as in the figure below.

2. Outline of the presentation/demo

We shall demo the Sketch Engine, a tool originally developed for dictionary-making
but now being re-engineered for language learners. Recent innovations include a
simplified interface, keyword lists, which allow users to find the words which are most

Optimizing the role of language in Technology-Enhanced Learning (2007): 21-23

22 A. KILGARRIFF

distinctive of a particular subcorpus, and ‘clustered word sketches’, word sketches in

which similar words are grouped together.
The word sketch is organized according to grammatical relations, with one list for
collocates in each different relation. The relation names (on blue backgrounds) head
each list. Collocates are listed according to the grammatical relation they occur in. In
contrast to summaries of commonly occurring near neighbours which do not apply
grammar, there is no junk: everything is there for an evident linguistic reason.
The first number is the actual number of occurrences of the collocation (taken from the
British National Corpus (BNC); all data used here is from the BNC.) The second
number is a salience statistic, used for sorting. When working online, the user can
click on the number and they are then shown the concordance for the collocation, so if
they are unsure what a word is doing in the word sketch, they can promptly find out.
Here, the items are lemmas (dictionary headwords) rather than word forms, so data for
goal and goals are merged. A ‘part of speech tagger’ has been applied to work out, for
example, where post is a verb (“post the letter”) and where a noun (“goal post”). The
word sketch as a whole is for the noun goal.
Word sketches were first used for the Macmillan English Dictionary for Advanced
Learners, and are now also being used at Oxford University Press, Collins, Chambers
Harrap, Le Robert and elsewhere. They changed the way the lexicographers used the
corpus. Rather than start with a KWIC concordance for the word, they went straight to
the word sketch, as that summarized most of what they needed the concordances for.

Word sketch for goal bnc freq = 10631 change options

and/or 1112 0.8 object_of 3430 3.1 subject_of 557 1.0 a_modifier 2546 1.8
objective 57 32.86 score 797 75.31 come 78 28.4 ultimate 83 42.22
try 30 32.67 achieve 363 48.14 give 34 14.57 away 25 32.56
goal 32 23.39 concede 126 47.79 win 13 14.32 winning 31 32.56
penalty 20 22.75 disallow 26 34.87 help 10 10.69 compact 34 31.79
target 22 20.1 pursue 75 33.13 stated 17 27.88
value 33 19.36 attain 34 29.34 adj_subject_of 149 1.4 late 53 27.33
conversion 12 18.92 net 18 26.7 important 10 15.32 dropped 11 26.98
aim 15 17.6 kick 36 26.2 organisational 22 26.83
mission 11 16.29 grab 30 24.43 long-term 34 25.7
priority 10 14.13 reach 78 23.81 common 56 24.62
strategy 11 12.28 set 97 23.53 headed 11 24.48
point 19 12.21 notch 10 22.81 organizational 18 24.45
 USING CORPORA IN LANGUAGE LEARNING: THE SKETCH ENGINE 23

n_modifier 1181 1.0 modifies 748 0.3 pp_after-p 58 7.1

drop 85 45.59 scorer 40 43.0 minute 37 39.18
penalty 100 45.27 difference 69 34.08
league 90 37.36 scoring 17 29.24 particle 86 4.5
consolation 24 35.39 Ace 18 28.33 back 32 28.93
opening 42 31.15 drought 14 26.56 down 32 28.62
second-half 13 30.46 Post 34 25.55 up 14 15.44
first-half 12 30.04 Kick 17 25.19
possessor 492 4.3
minute 30 21.09 keeper 16 24.71
England 12 13.95
half 17 19.15 weight 21 21.01
policy 42 18.73 Lead 16 20.29
pp_from-p 275 4.1
relationship 16 13.36 average 10 17.56
attempt 12 17.09
development 22 13.22 setting 11 16.98

Goals occur, of course, in sport as well as life. The word sketch highlights the
ambiguity. Scanning the ‘object-of’ list, if we score, concede, disallow, net or kick
goals, we are talking sport; if we achieve, pursue, attain or reach them, life. England
football fans will be glad to see England standing alone in the ‘possessor’ relation to
goals!
Word sketches can be explored at http://www.sketchengine.co.uk where papers and
bibliographical references are also available.

3. Issues and challenges

The challenge is to establish the case for corpora in language teaching in general. The
second is the attractiveness and usability of the SkE for language learners. The third is
of finding or developing appropriate corpora for language learners to use.

4. Reference
Kilgarriff A., Rychly P., Smrz P. and Tugwell D. (2004), “The Sketch Engine”, in
G. Williams and S. Vessier (eds), Proceedings of the Eleventh EURALEX International
Congress, EURALEX 2004, Lorient, France, July 6-10, 2004. Lorient: Faculté des Lettres
et des Sciences Humaines, Université de Bretagne Sud (Proc EURALEX 2004), Lorient.
 A plurilingual ICALL System for Romance
languages

Thomas Koller
University of Nottingham

1. Introduction
The (completed Ph.D.) research described in this abstract deals with the design,
development, implementation and evaluation of an interactive plurilingual ICALL
(Intelligent Computer-Assisted Language Learning) software system (ESPRIT) for
contrastive learning of French, Spanish and Italian. ESPRIT targets learners who are
already at an advanced level in at least one of the Romance languages involved. These
learners are expected to be familiar with general lexical and grammatical properties of
this language. Equivalent properties of the other languages are taught through
comparison.
The addressed research questions build upon the general research findings in
plurilingual teaching and learning of Romance languages, CALL (Computer-Assisted
Language Learning) and ICALL, and the use of animation in language teaching.
Formative and summative evaluation processes provided learner assessment data of
different components of ESPRIT.
Plurilingual means that grammatical and lexical properties of the languages involved
are tightly linked to each other, showing a high degree of similarity in form and
function. Plurilingual teaching and learning of Romance languages exploits the
similarities between these languages to teach them contrastively and to raise the
language awareness of the learner.
The ESPRIT toolset comprises dictionary tools, a concordancer, an input analysis and
feedback module, custom-made animated grammar presentations and an authoring tool
for animated text. ESPRIT represents a fully functional web-based language learning
platform which is designed for autonomous learning. ESPRIT uses a TV metaphor to
present language learning materials to the learner. The contents can easily be expanded
at any time.
In ESPRIT, learners are free to explore the activities offered and to choose the
activities which are of most interest to them. Guided tours, however, provide
information and help about which activities form a logical unit, and can be used to
suggest in which sequence to work on materials.

Optimizing the role of language in Technology-Enhanced Learning (2007): 25-27

26 TH. KOLLER

2. Outline of the presentation/demo

After providing a short introduction on plurilingual teaching and learning, my
presentation will focus on the demonstration of the toolset and the web-based language
learning platform developed for ESPRIT. Special emphasis will be put on (a) the NLP
techniques integrated into several tools and (b) the applicability of ESPRIT tools and
resources to other projects.
Although the similarities between Romance languages have been described
extensively in contrastive linguistics for decades, a broader interest in research on
plurilingual teaching and learning only emerged in the 1990s. Since then, several
European projects have been devoted to plurilingual teaching and learning of Romance
languages. The materials developed in these projects do not involve Natural Language
Processing (NLP) capabilities and almost exclusively focus on receptive skills.
Plurilingual teaching is potentially highly effective, yet plurilingual teaching and
learning material is quite hard to obtain. Existing materials only contain a limited
amount of reading texts and exercises.
The innovative character of my Ph.D. research lies in the investigation of NLP
techniques to enhance the plurilingual teaching and learning of Romance languages. In
contrast to existing plurilingual materials, ESPRIT tools allow the learner to work
independently on unrestricted learner-retrieved text and to obtain dynamic feedback on
learner input. I aimed to develop flexible and interactive, easily expandable software
which supports plurilingual teaching and learning of Romance languages and which
helps language learners to optimally exploit their existing knowledge in any one
Romance language. The single tools and the web-based language learning platform
developed for ESPRIT are available at any time on the Internet.

Tools and language data of ESPRIT have already been reused in current projects. Due
to their modular character, the tools and language data can easily be integrated in any
other project, in which they can be applied to other languages or even language
families. Slavic languages, for example, also share a high number of grammatical and
lexical properties.
Several ESPRIT tools can also be adapted and provided as Firefox browser plug-ins.
As a Firefox extension, an ESPRIT tool would be instantly accessible from any other
web page (for example for dictionary look-up). The dictionary tools, lexicon interface
components and the concordancer could be adapted as Firefox extensions to provide a
wide range of plug-in resources for plurilingual learning of Romance or other
languages.

3. Issues and challenges

When learning a third or any further language, learners automatically create links
between the properties of currently learned and already learned languages. The
plurilingual teaching and learning method, however, is largely unknown to learners
 A PLURILINGUAL ICALL SYSTEM FOR ROMANCE LANGUAGES 27

(and teachers). The evaluation for ESPRIT also showed that adult learners varied
considerably in the number and type of languages learned already and the degree of
fluency therein. Therefore the development of materials for plurilingual teaching and
learning posed a number of issues and challenges which differ from second language
acquisition and the creation of monolingual language learning materials.
Foreign language teaching in secondary schools and at universities has been largely
unaffected by plurilingual research. Language students at both levels only occasionally
get the opportunity to learn similar languages simultaneously in a plurilingual setting.
As a consequence, it is challenging to identify target learners and to conduct standard
institutionalised testing and evaluation of developed plurilingual materials.
The development of plurilingual materials in general has in many cases not been
directly connected to research in third language acquisition. Additionally, the majority
of existing plurilingual materials tends to be rather descriptive than didactic.
Therefore, in my opinion, it would be beneficial for future research in plurilingual
teaching and learning to be more tightly linked to research findings of third language
acquisition.

4. References
BLANCHE-BENVENISTE C. (ed.) (1997), EuRom 4: Metodo de ensino simultâneo das línguas
românicas – Metodo para la enseñanza simultánea de las lenguas románicas – Metodo di
insegnamento simultaneo delle lingue romanze – Méthode d’enseignement simultané des
langues romanes, Florence, La Nuova Italia Editrice.
DEGACHE Christian (ed.) (2003), Intercompréhension en langues romanes. Du développement
des compétences de compréhension aux interactions plurilingues, de Galatea à Galanet,
volume 28. Grenoble, LIDILEM, Université Stendhal Grenoble 3.
KOLLER Th. (2007), Design, Development, Implementation and Evaluation of a Plurilingual
ICALL System for Romance Languages Aimed at Advanced Learners, PhD thesis, Dublin
City University, Dublin.
MCCANN W.J., KLEIN H.G. and STEGMANN T.D. (2002), EuroComRom – The Seven Sieves,
volume 5 of Editiones EuroCom. Aachen, Shaker.
SCHMIDELY J., ALVAR EZQUERRA M. and HERNÁNDEZ GONZÁLEZ C. (2001), De una a cuatro
lenguas. Intercomprensión románica: del español al portugués, al italiano y al francés.
Madrid, Arco Libros.
 Like stars in the firmament:
language learning on mobile devices

Agnes Kukulska-Hulme
The Open University, UK

1. Introduction
Mobile learning can be studied as one instance of the ongoing adoption of innovative
technologies in the field of education, particularly with a view to understanding learner
experience and the potential of the new technologies to transform current practices.
Educational uses of mobile technologies offer a rich and complex field of investigation
which allows me personally to combine my expertise in e-learning pedagogy with my
background in linguistics, language learning, dictionary design and terminology
studies (areas I was actively involved in during the 1980s/90s). I’m particularly
interested in how mobile devices are changing foreign language learning and how new
forms and motivations for language learning might in turn have an effect on attitudes
and approaches to multilingual knowledge seeking, global communication and
knowledge representation on the web.
My research in mobile learning has been fairly wide-ranging, encompassing studies of
how learners read course materials on mobile devices (Waycott and Kukulska-Hulme
2003), surveys of learner-driven mobile innovation (Kukulska-Hulme and Pettit 2006;
Pettit and Kukulska-Hulme 2007), critical reviews of evaluation in mobile learning
(Traxler and Kukulska-Hulme 2006), reflections on what has been learnt with regard
to mobile device usability (Kukulska-Hulme 2007), and issues of collaboration and
privacy in contextual learning (Kukulska-Hulme et al. 2007). Together with my
colleague John Traxler I co-edited the first book on mobile learning to give a coherent
account of the field, incorporating a dozen international case studies (Kukulska-Hulme
and Traxler 2005). My externally funded projects have also led to the publication of a
guide to innovative e-learning with mobile technologies, distributed widely within UK
higher and further education. I have tried to make sense of how the field is evolving by
studying the possibilities of both formally-designed and user-driven mobile learning
(Kukulska-Hulme, Traxler and Pettit 2007). I have also attempted to imagine how
mobile language learning will develop (Kukulska-Hulme 2006; Kukulska-Hulme
forthcoming).

Optimizing the role of language in Technology-Enhanced Learning (2007): 29-31

30 A. KUKULSKA-HULME

2. Outline of the presentation

Mobile learning is a fast-moving field. It is becoming clearer that device ownership is
a factor in adoption, in the type of activity that learners are likely to engage in, and in
the integration of learning activities with other aspects of daily life. However, many
aspects of mobile learning remain under-explored; for example, connections between
mobile and online activity have not yet been investigated in any systematic way, and
neither have the implications for the ways that language communication and textual
content are used or represented on mobile devices. One interesting observation is that
mobile learning projects and initiatives are likely to have outcomes that had not been
anticipated by educators or providers of language learning materials, e.g. Gilgen
(2005) reports that “several new and unexpected uses and results” (p.32) came about in
their mobile learning projects.
In my presentation, I will first share and discuss a working classification I have
developed of models of participation in mobile language learning: Institutional
adoption model; Content delivery model; Proposed activity model; Specified activity
model; Content sharing model; and User-generated activity model (Kukulska-Hulme,
forthcoming). I’d like to consider the implications for the types of conversations that
learners are able to have in these different models of interaction and the extent to
which they are able to take part as initiators of learning activities or contributors of
language material. If we consider for a moment that each mobile user is interconnected
with others, to what extent are they able to be noticed as shining stars in a
metaphorical firmament (alluded to in the title of my presentation)? What prospects
are there for mobile devices to give learners new opportunities to observe how
language is used, request specific types of language support, or share their findings
and ideas with others?

3. Issues and challenges

Language captured and shared in context: Mobile devices are well known for
facilitating learning in context – bringing learning closer to real life situations, either
spontaneously or in environments designed for those ends. Learners can gather
primary data on location, for example by capturing instances of language in use, or use
their device to capture and share reflections on language problems and needs, the
moment they arise. What are the best ways of fulfilling this potential?
Fragmented conversations: Mobile interaction can result in a fragmented experience,
especially when use of an online forum is also part of the learning design. We know
very little about the learning conversations that mobile devices can, and cannot,
facilitate. Ethical and practical issues get in the way of analysing interactions; the
specific constraints around mobile learning are still poorly understood.
Usability: Often the first thing that people remark on, when they consider mobile
learning, is the difficulty of reading from a small screen on a mobile device, or how
hard it might be to construct appropriate short text messages within the context of
 LIKE STARS IN THE FIRMAMENT: LANGUAGE LEARNING ON MOBILE DEVICES 31

education. Some then begin to see these difficulties as opportunities or challenges –

perhaps shorter texts are better, and have educational value, such as training students
to summarize their thoughts or getting teachers to be more precise about instructions.
I’m interested in how issues of usability can act in a positive way to instigate
reflections on educational goals and changing literacies.

4. References
CHINNERY G.M. (2006), “Going to the MALL: Mobile Assisted Language Learning”, in
Language Learning & Technology, 10(1): 9-16. Available online: http://llt.msu.edu/
vol10num1/emerging/default.html
KUKULSKA-HULME A. and TRAXLER J. (eds) (2005), Mobile Learning: A Handbook for
Educators and Trainers, Routledge, London.
KUKULSKA-HULME A. (2006), “Mobile Language Learning Now and in the Future”, in
Svensson, P. (ed.), Från vision till praktik: Språkutbildning och Informationsteknik (From
vision to practice: language learning and IT), Swedish Net University (Nätuniversitetet):
295-310.
KUKULSKA-HULME A. (2007), “Mobile usability in educational contexts: what have we
learnt?”, Special issue of the The International Review of Research in Open and Distance
Learning, 8(2): 1-16. Available online: http://www.irrodl.org/index.php/irrodl
KUKULSKA-HULME A., Traxler J. and Pettit J. (2007), “Designed and User-generated Activity
in the Mobile Age”, in Journal of Learning Design, 2 (1): 52-65. Available online:
http://www.jld.qut.edu.au/
PETTIT J. and KUKULSKA-HULME A. (2007,)” Going with the Grain: Mobile Devices in
Practice”, in Australasian Journal of Educational Technology (AJET), 23 (1): 17-33.
Available online: http://www.ascilite.org.au/ajet/ajet23/ajet23.html
SHARPLES M. (2006), Big Issues in Mobile Learning. Kaleidoscope report, Available online:
http://mlearning.noe-kaleidoscope.org/repository/BigIssues.pdf
WAYCOTT J. and KUKULSKA-HULME A. (2003), “Students’ Experiences with PDAs for
Reading Course Materials”, in Personal and Ubiquitous Computing, 7 (1): 30-43.
 Computer-mediated communication
for language learning

Marie-Noëlle LAMY
The Open University, UK

1. Introduction
The field of activity captured by the phrase “computer-mediated communication for
language learning” recently reached a critical mass, as regards the number of teaching
projects and of published papers that have been devoted to it. Chun (2007) suggests
that ‘communication’ as used in the phrase ‘computer-assisted communication’
(CMC) receives the most coverage of all topic categories in her overview of recent
research based on evidence from two major US journals on technology-mediated
language learning, and also comes top of a list of ‘hits’ tracked by one of the two
journals in her corpus.
Although caveats are needed due to exclusively US-oriented nature of these results,
similar trends are observed in other research cultures, signalling that CMC for
language learning (henceforth CMCL) as a field is no longer immature and can be held
up to scrutiny. In a volume to be published in November 2007, Lamy and Hampel
offer such a scrutiny. The current presentation gives a preview of their findings, and
outlines research directions suggested not only by the gaps identified in their study but
also by the emergence of new questions raised within neighbouring areas such as
multiliteracies research.

2. Outline of the presentation

In this section, I present a brief overview of activity in the practice and research of
CMCL. I start with the methodological relationships that can be established between
CMCL and three related fields, which are: generic (i.e. non-language-oriented)
educational CMC, socio-personal CMC, and Computer-Assisted Language Learning
or CALL. In the latter field, Warschauer (1995) put CMCL on the map by publishing
the first practitioner book on the topic. According to him, the hopes of early adopters
of CMCL included giving learners the opportunity to:
• communicate with native speakers and with each other either one-to-one or, more
innovatively, one-to-many and many-to-many;

Optimizing the role of language in Technology-Enhanced Learning (2007): 33-35

34 M.-N. LAMY

• plan their communication;

• revisit their work, owing to the permanent traces made available to them through
the technologies.
I assess whether these expectations been met, and I identify new questions that have
arisen along the way, through interrogating the ERIC (Educational Resources
Information Centre) database for the years 1992−2005. The results shows that
although Warschauer was right to predict a boom in remote communication for
language teaching, the hopes of early adopters, as he listed them, were not all fulfilled.
Furthermore, a number of meta-studies published since the mid-90s, as well as
Warschauer and Kern’s own (2000) review, have shown that the “simple question to
which everyone wants an answer ‘Does the use of network-based language teaching
lead to better language learning?’ […] turns out to be not so simple”, and that the
CMCL community might do better to abandon the search for improvements in
language learning and instead “look to particular practices of use including the
specifics of learner profiles, task types, process description, discourse, interaction
patterns and formal outcomes” (Warschauer and Kern, 2000: 2; original emphasis). By
reference to 7 meta-studies, I identify these practices of use and I offer a ‘health check’
of the field, highlighting its achievements, but also the over-coverage of certain topics
(e.g. student participation patterns), the under-coverage of others (e.g. assessment) and
the recurring concerns expressed by meta-study authors about the quality of research
in the CMCL literature.

3. Issues and challenges

In this section of the presentation, I identify three types of challenge associated with CMCL.
The first one, the unsatisfactory functioning of the practice-research feedback loop, is
not specific to this field but can be seen with particularly sharpness in CMCL, perhaps
because of the spectacular expansion of the field in a relatively short time.
Major challenge number 2 is at the level of pedagogical practice. I focus on tensions
that have not had much exposure in the CMCL research literature so far. Of the two
broad groups of tools, text-based and voice-based environments, I argue that both
suffer from insufficient attention to the machine-mediated nature of the activity. Many
practitioners now agree that online language classes have specific socio-affective or
intercultural needs (as CMCL research has indeed been showing) but teachers remain
vulnerable to learner disengagement through confusion and overload, because the
materiality of the environments is taken as a given rather than being made an explicitly
part of the pedagogical considerations that inform teaching design.
Major challenge number 3 is a theoretical and methodological one for researchers. It
relates the the multimodal nature of electronic environments and it concerns the choice
of theoretical and methodological frameworks for the analysis of learner
conversations, when such conversations may be carried out via a range of interrelating
semiotic systems, some wired into the machine and others freely deployed or created
by the learners, many of them of a non-linguistic nature.
 COMPUTER-MEDIATED COMMUNICATION FOR LANGUAGE LEARNING 35

Finally, I show how the nature of challenges 2 and 3 point to the importance of
multiliteracies research for the future of CMCL.

4. References
CHUN D.M. (2007), “Come Ride the Wave: But Where is it Taking Us?”, in The CALICO
Journal 24(2): 239-252.
HASSAN X., HAUGER D., NYE G. and SMITH P. (2005), “The Use and Effectiveness of
Synchronous Audiographic Conferencing in Modern Language Teaching and Learning
(Online Language Tuition): A Systematic Review of Available Research”, in Research
Evidence in Education Library, London: EPPI-Centre, Social Science Research Unit,
Institute of Education, University of London.
HUBBARD P. (2005), “A Review of Characteristics in CALL Research”, in Computer Assisted
Language Learning 18(5): 351-368.
JUNG U. (2005), “CALL – Past, Present and Future: A Bibliometric Approach”, in ReCALL
17(1): 4-17.
KERN R. G. (2006), « La Communication médiatisée par ordinateur en langues: recherches et
applications récentes aux USA », in F. Mangenot and C. Dejean-Thircuir (eds), Les
Echanges en ligne dans l’apprentissage et la formation, le français dans le monde,
recherches et applications 40: 17-29.
LEVY M. (2000), “Scope, Goals and Methods in CALL Research: Questions of Coherence
and Autonomy”, in ReCALL 12(2): 170-195.
LIU M., MOORE Z., GRAHAM and LEE, S. (2002), “A Look at the Research on Computer-
based Technology Use in Second-language Learning: A Review of the Literature from
1990-2000”, in Journal of Research on Technology in Education 34(3): 250-273.
LAMY M.-N. and HAMPEL R. (2007 forthcoming)? Online Communication for Language
Teaching and Learning, Palgrave Basingstoke, Palgrave McMillan.
WARSCHAUER M. (1995), Virtual Connections: Online Activities and Projects for Networking
Language Learners, Honolulu: Second Language Teaching and Curriculum Centre,
University of Hawaii.
WARSCHAUER M. and KERN R. (eds) (2000), Network-based Language Teaching: Concepts
and Practice, Cambridge, Cambridge University Press.
ZHAO Y. (2003), “Recent Developments in Technology and Language Learning: A Literature
Review and Meta-analysis”, in The CALICO Journal 21(1): 7-27.
 Writing English as a second language:
A proofreading tool

Claudia Leacock
The Butler Hill Group

1. Introduction
This work combines Natural Language Processing (NLP) and Machine Learning (ML)
techniques for detecting and correcting grammatical errors in the writing of English
Language Learners (ELL).
The Writing English as a Second Language tool being developed at Microsoft
Research (for which the presenter is a consultant) focuses on those areas of grammar
that pose special challenges for English language learners. This presentation focuses
on those problems that are hardest for ELLs – the use of determiners and of
prepositions – although the system also identifies gerund/infinitive confusion,
auxiliary verb presence and choice, over-regularized verb inflection (writed vs. wrote),
adjective/noun confusion (China book vs. Chinese book), word order errors, and mass
vs. count noun errors (much knowledge vs. many knowledges).

2. Outline of presentation
I will describe the system’s major components:
1. The Suggestion Provider consists of Individual error identification modules that
identify potential errors. These modules are flexible and can identify errors using
ML techniques rules, regular expressions or a combination of the three.
For the preposition and determiner correction modules, a classifier is used that is
trained on edited native English. For each potential insertion point of a determiner
or preposition in that training data, a vector of features is extracted from the
context.
2. The Language Model, which selects the most likely suggestion(s), is a 5-gram
model trained on the English gigaword corpus.
3. The Example Provider retrieves relevant example sentences from the web to help
the user select the most appropriate rewrite. This innovative component generates

Optimizing the role of language in Technology-Enhanced Learning (2007): 37-39

 C. LEACOCK 38

an exact string query including a window of context around the suggested

correction. The query is issued to a search engine, and the retrieved sentences are
ranked and presented to the user.
For system accuracy, we will present two different evaluations: (1) Automatic
evaluation on copyedited native text – under the assumption that it contains no errors.
(2) Human evaluation of essays written by Chinese ELLs.

3. Issues and challenges

Given the very high frequency with which prepositions and determiners occur in
English, the false flag rate must be very low in order to be acceptable. While a native
speaker can easily identify and ignore a false flag, language learners would have to
take time inspecting the example sentences to decide which is correct – and even then
may get confused. The system currently uses handcrafted heuristics to minimize false
flags – which is laborious and requires retuning each time a model is retrained. The
system’s developers are investigating a learned ranker to replace the handcrafted
heuristics.
Traditionally, grammar checkers give the writer a mini-multiple-choice test when
providing potential corrections. This is fine for native writers – but simply poses
another challenge for language learners. The Example Provider is an innovative
method for enabling the user to make an informed decision. Initial response has been
positive but will require refinements based on the results of extensive user testing.

4. References
BURSTEIN J. and LEACOCK C. (eds) (2006), Natural Language Engineering: Special Issue on
Using NLP in Educational Applications, 12:2.
BURSTEIN J., CHODOROW M. and LEACOCK C. (2004), “Automated Essay Evaluation: The
Criterion Online Writing Service”, in AI Magazine 25:3: 27-36.
BURSTEIN J., CHODOROW M. and LEACOCK C. (2003), “Criterion Online Essay Evaluation:
An Application for Automated Evaluation of Student Essays”, in Proceedings of the
Fifteenth Annual Conference on Innovative Applications of Artificial Intelligence
Conference (IAAI-03), Acapulco.
CHODOROW M., TETREAULT J.R. and HAN N.-R. (2007), “Detection of Grammatical Errors
Involving Prepositions”, in Proceedings of the 4th ACL-SIGSEM Workshp on Prepositions:
25-30.
CHODOROW M. and LEACOCK C. (2000), “An Unsupervised Method for Detecting
Grammatical Errors”, in Proceedings of the 1st Annual Meeting of the North American
Chapter of the Association for Computational Linguistics, Seattle, WA.
HAN N.-R., CHODOROW M. and LEACOCK C. (2006). “Detecting Errors in English Article
Usage by Non-Native Speakers », in Natural Language Engineering 12:2.
HAN N.-R., CHODOROW M. and LEACOCK C. (2004). “Detecting Errors in English Article
Usage with a Maximum Entropy Classifier Trained on a Large, Diverse Corpus”, in
Proceedings of the 4th International Conference on Language Resources and Evaluation.
Lisbon.
 WRITING ENGLISH AS A SECOND LANGUAGE: A PROOFREADING TOOL 39

IZUMI E., UCHIMOTO K., SAIGA T., SUPNITHI T. and ISAHARA H. (2003), “Automatic Error
Detection in the Japanese Learners’ English spoken Data”, in The Companion Volume to
the Proceedings of 41st Annual Meeting of the Association for Computational Linguistics,
July 2003: 145-148.
LEACOCK C., and CHODOROW M. (2003), “Automated Grammatical Error Detection”, in M.D.
Shermis and J. Burstein (eds), Automated Essay Scoring: A Cross-Disciplinary
Perspective. Hillsdale, NJ, Lawrence Erlbaum : 195-208.
LEACOCK C., and CHODOROW M. (2001), “A Corpus-Based Approach to Diagnosing
Grammatical Errors”, in Corpus Linguistics and Language Teaching Conference, Boston,
MA.
LIU T., ZHOU M., GAO J., XUN E. and HYAN C. (2000), “PENS: A Machine-Aided English
Writing System for Chinese users”, in Proceedings of ACL 2000: 529-536.
SHERMIS M.D., BURSTEIN J. and LEACOCK C. (2006), “Applications of Computers in
Assessment and Analysis of Writing”, in C.A. McArthur, S. Graham and J. Fitzgerald
(eds), Handbook of Writing Research, New York, Guilford Press.
TURNER J. and CHARNIAK E. (2007), “Language Modeling for Determiner Selection”, in
Human Language Technologies 2007: The Conference of the NAACL; Companion
Volume, Short Papers: 177-180.
 Second language acquisition theory and TELL

Fanny Meunier
Université Catholique de Louvain

1. Introduction
As a researcher in Second Language Acquisition (and more specifically instructed
second language acquisition) and a teacher of English as a foreign language, my aim in
this presentation is twofold: first, stress the importance of the two ‘L’s in technology-
enhanced language learning; and secondly, address the convergences and divergences
that exist between big issues in second language acquisition and TELL.

2. Outline of the presentation

The appeal of new technologies in language learning has undoubtedly played a role in
downgrading the focus on teaching/learning methodologies per se, be they technology-
enhanced or not. Wible (2005:2) even states that whilst massive resources are invested
in the development of information technology for e-learning, little concentrated effort
is devoted to bridging the gap between technology and second/foreign language
education. I will argue that an additional gap still increases the complexity of the
situation, i.e. the gap that exists between SLA theory and second/foreign language
education.
According to Wible again (2005:72), TELL would benefit from a shift from what the
technology is capable of doing to what the learner actually needs. I will demonstrate in my
presentation that such a shift ideally requires insights into SLA theory.
The issues that will be dealt with include the positive or negative influence that
technology may have in addressing some internal/external variables in SLA (L1
background, learners’ characteristics, learning styles, input, etc.), the
receptive/productive skills dichotomy, a number of cognitive aspects (awareness,
saliency, elaboration, rehearsal, etc.) and types of feedback.

3. Issues and challenges

First, closer collaboration should be encouraged between the technological,
acquisitional and methodological paradigms of language learning. Secondly, TELL
should consider what I call (see Meunier forthcoming) ‘principled eclecticism in

Optimizing the role of language in Technology-Enhanced Learning (2007): 41-42

42 F. MEUNIER

learning’ as one of its future challenges. Learners in TELL environments should have
access to observational, descriptive or explanatory options, together with opportunities
for immediate feedback. Third, learnability issues (defined here as the input/output
efficiency of some method or approach) should become more central in order to
validate the efficiency of TELL.

4. References
CHAPELLE C. (1998), “Multimedia CALL: Lessons to be Learned from Research on
Instructed SLA”, in Language Learning and Technology, 2(1): 22-34.
CHAPELLE C. (2003), English Language Learning and Technology, Benjamins, Amsterdam
and Philadelphia.
KASPER L. (2000), “New technologies, new literacies: focus discipline research and ESL
learning communities”, in Language Learning & Technology, vol. 4, n°. 2: 105-128.
LIGHTBOWN P. and SPADA N. (2003) Factors Affecting Second Language Learning. How
Languages Are Learned, Revised edition, Oxford University Press, Oxford.
MEUNIER, F. (forthcoming) “Corpora, cognition and pedagogical grammars: An account of
convergences and divergences”, in S. De Knop and T. De Rycker (eds), Cognitive
Approaches to Pedagogical Grammar, Mouton de Gruyter, Berlin.
WIBLE D. (2005), Language Learning and Language Technology: Toward Foundations for
Interdisciplinary Collaboration, Crane, Taipei.
WIBLE D. (in press), “Multiword Expressions and the Digital Turn”, in F. Meunier and
S. Granger (eds), Phraseology in Foreign Language Learning and Teaching, John
Benjamins, Amsterdam.
 Detecting syntactic interference

John Nerbonne
University of Groningen

Introduction
This presentation involves joint work with Wybo Wiersma (Groningen) and Timo
Lauttamus (Oulu). It applies techniques from quantitative computational linguistics to
the problem of detecting frequent effects of first language interference in second
language learning. We focus on production interference in syntax.
Second language learners typically differ syntactically from native speakers not only
in making outright errors, but also in overusing and under-using some constructions –
all of which we subsume under INTERFERENCE. In approaching the phenomenon of
interference computationally, we were motivated both to attempt to identify
interference effects more systematically, and also to attempt to quantify a level of
aggregate interference, a goal Weinreich (1953: 63) found worthwhile, but which he
speculated to be unreachable:
No easy way of measuring or characterizing the total impact of one language on another in
the speech of bilinguals has been, or probably can be devised. The only possible
procedure is to describe the various forms of interference and to tabulate their
frequency.

2. Outline
Following a suggestion by Aarts and Granger (1998), we model the syntax of the
second-language learners of English via the parts of speech (POS) they use and the
sequences in which the POS appear. The idea is to compare the POS sequences used
by second-language learners to those used by natives. Concretely, we examine the
distribution of triplets of POS in a large corpus of English as used by adult Finnish
immigrants to Australia, and compare this distribution to that of their children, who
immigrated as children and speak English at a near-native level. To assay the “total
impact” as Weinreich wished, we examine the differences between the two
distributions via a permutation test, which is implemented in a Monte Carlo fashion.
To identify systematically the areas of difference, we examine the frequent POS triples
that contribute most to the overall differences in the two distributions.

Optimizing the role of language in Technology-Enhanced Learning (2007): 43-44

 J. NERBONNE 44

3. Issues and challenges

Although we can assay “total impact” and also identify areas of syntactic differences
straightforwardly, there are several points at which improvement would be useful and
interesting. First, the technique works at a high level of aggregation. This is not a
serious problem in language contact study, which is interested in exactly such
population effects, but it is a problem if one wishes to analyse the work of individual
second-language learners. Second, and related, we should wish to study the influence
of individual speakers on the approach, as Sanders (2007) has. Third, our approach
assumes that POS sequences represent syntax well (for the purpose of assaying
differences), an assumption which is justified by the endocentricity of syntax, but this
assumption could also be tested. Finally, we need to analyse more data sets involving
different languages since the Finnish effects on English may be a special case.
We are also interested in receptive interference (Moberg et al. 2007), but will not have
the time to present that work here.

4. References
AARTS J. and GRANGER S. (1998), “Tag sequences in learner corpora: A Key to Interlanguage
Grammar and Discourse”, in S. Granger (ed), Learner English on Computer, London,
Longman: 132-141.
MOBERG J., GOOSKENS C., NERBONNE J. and VAILLETTE N. (ca. 2007), “Conditional Entropy
Measures Intelligibility among Related Languages”, accepted to appear in F. Van Eynde,
P. Dirix, I. Schuurman and V. Vandeghinste (eds.) Proceedings of Computational
Linguistics in the Netherlands 2006, Amsterdam, Rodopi.
NERBONNE J. and WIERSMA W. (2006), “A Measure of Aggregate Syntactic Distance”, in
J. Nerbonne and E. Hinrichs (eds), Linguistic Distances, Workshop at the joint conference
of International Committee on Computational Linguistics and the Association for
Computational Linguistics, Sydney, July, 2006: 82-90.
SANDERS N. (2007) “Measuring Syntactic Difference in British English”, in Proceedings of
the ACL 2007 Student Research Workshop, Prague: ACL: 1-6.
WEINREICH, Uriel ([1953], 1968, 1974), Languages in Contact. The Hague, Mouton.
 Planning a smart phone system to support
self-directed L2 vocabulary learning

Richard Pemberton
University of Nottingham

1. Introduction
There are two major problems with vocabulary learning that almost every language
learner will be familiar with:
• learning enough frequent vocabulary to be able to read and listen fluently;
• retaining the vocabulary that we have learned.

The first problem involves a considerable amount of time. To take English for
example, in order to be able to understand unsimplified texts, you need to know some
3,000–4,000 of the most common English word families (Nation & Waring 1997;
Nation 2001). The figure is likely to be upwards of 5,000 word families if fluent
reading for pleasure is the aim.
Equally, if vocabulary is to be retained, the learner needs to spend a lot of time in
conscious processing or repeating of the target items (explicit learning) and/or in
extensive language use (implicit learning). These problems of time are of course even
worse for the busy adult learner living outside the target country.
One type of technology which has the potential to save time on the go and to support
both implicit and explicit learning is the mobile phone. However, recent mobile phone
systems supporting vocabulary learning have tended to use one medium only – e.g.
text messages (Pincas 2004; Song & Fox 2005), e-mail (Thornton & Houser 2005) or
images (Joseph et al 2005) – and to have involved designed rather than learner-located,
learner-generated and learner-shared materials and activities.

2. Outline of presentation
In this presentation I will first propose a smart phone system that could use the
phone’s full capabilities (see e.g. Kukulska-Hulme & Shield 2007: 20) to support self-
directed vocabulary learning.
I will then exemplify and discuss the desirability of various potential features (both
‘existing’ and ‘to-be-created’), including:

Optimizing the role of language in Technology-Enhanced Learning (2007): 45-47

46 R. PEMBERTON

• the creation of subtitles for video recordings (cf. Fallahkhair et al 2007);

• OCR scanning of camera shots of written text;
• the use of a personalisable wordlist/dictionary with testing functions;
• a testing ‘scheduler’ based on the principles of spaced repetition and expanding
rehearsal (Ellis 1995);
• the use of a ‘producing’ dictionary with associative functions;
• geotagging;
• a simple advising system to provide guidance re vocabulary learning.

3. Issues and challenges

There are a number of issues to be clarified and challenges to be met, including:
• How long will it take before more phone models provide enough screen size for
subtitled videos?
• Would it be too expensive for Europe? Which countries would it be best to target?

4. References
ELLIS N.C. (1995), “The psychology of foreign language vocabulary acquisition: implications
for CALL”, in Computer Assisted Language Learning 8(2-3): 103-128.
FALLAHKHAIR S., PEMBERTON L. and GRIFFITHS R. (2007), “Development of a cross-platform
ubiquitous language learning service via mobile phone and interactive television”, in
Journal of Computer Assisted Learning 23: 312-325.
JOSEPH S., BINSTED K. and SUTHERS D. (2005), “PhotoStudy: vocabulary learning and
collaboration on fixed and mobile devises”, in H. Ogata, M. Sharples, G. Kinshuk and Y.
Yano (eds), Proceedings of the third IEEE International Workshop on Wireless and
Mobile Technologies in Education 2005, Los Alamitos, CA: IEEE: 206-210.
KUKULSKA-HULME, A. and SHIELD L. (2007), “An overview of mobile assisted language
learning: can mobile devices support collaborative practice in speaking and listening?”, in
EuroCALL 2007. http://vsportal2007.googlepages.com/collaborativepractice [Accessed 15
September 2007]
KUKULSKA-HULME A., TRAXLER J. and PETTIT J. (2007), “Designed and user-generated
activity in the mobile age”, in Journal of Learning Design 2(1): 52-65.
http://www.jld.qut.edu.au [Accessed 15 September 2007.]
NATION I.S.P. (2001), Learning Vocabulary in Another Language, Cambridge, Cambridge
University Press.
NATION P. and WARING R. (1997), “Vocabulary size, text coverage and word lists », in
N. Schmitt and M. McCarthy (eds), Vocabulary: description, acquisition and pedagogy,
Cambridge, Cambridge University Press: 6-19.
PINCAS A. (2004), “Using mobile phone support for use of Greek during the Olympic Games
2004”, in International Journal of Instructional Technology & Distance Learning
http://www.itdl.org/Journal/Jun_04/article01.htm [Accessed 19 September 2007.]
SONG Y. and FOX R. (2005), “Integrating web-based ESL vocabulary learning for working
adult learners”, in H.Ogata, M. Sharples, G. Kinshuk and Y. Yano (eds), Proceedings of
the third IEEE International Workshop on Wireless and Mobile Technologies in Education
2005, Los Alamitos, CA, IEEE: 154-158.
THORNTON P. and HOUSER C. (2005), “Using mobile phones in English education in Japan”,
in Journal of Computer Assisted Learning 21: 217-228.
 SMART PHONE SYSTEM TO SUPPORT SELF-DIRECTED L2 VOCABULARY LEARNING 47

THORNTON P. and SHARPLES M. (2005), “Patterns of technology use in self-directed Japanese

language learning projects and implications for new mobile support tools”. In H.Ogata, M.
Sharples, G. Kinshuk and Y. Yano (eds), Proceedings of the third IEEE International
Workshop on Wireless and Mobile Technologies in Education 2005, Los Alamitos, CA,
IEEE: 203-205.
 The dictionary of the future

Michael Rundell
Lexicography MasterClass Ltd and Macmillan Dictionaries

1. Introduction
Donald Rumsfeld’s famous reflections on “what we know we don’t know and what we
don’t know we don’t know” apply to most forms of futurology, and certainly to any
attempt to predict what might happen in the world of reference materials. This talk is
at the interface of language-learning, lexicography, NLP, and delivery media, and will
outline some possible future directions for dictionaries aimed at learners of English.

2. Outline of the presentation/demo

The monolingual learner’s dictionary (MLD) follows a model whose essential
characteristics were developed in the 1930s (by people like Harold Palmer and
Michael West) and found concrete form in A.S Hornby’s ur-MLD, published in 1942.
This was later to morph into the Oxford Advanced Learner’s Dictionary, which is still
going strong after 60 years. There have been two big changes in the intervening
period:

the arrival of corpora in the 1980s, which led to great improvements in quality
as lexicographers got access to objective language data (Sinclair 1987);

the growth of competition: Hornby’s dictionary had the field to itself till 1978,
but four other contenders have since entered the fray, and this has helped to
drive innovation (Rundell 1998).

Though content and accessibility of these dictionaries has steadily improved, the basic
model hasn’t fundamentally altered. But like any other kind of reference resource, the
MLD can’t fail to be affected by the biggest change of all – the arrival of the Web.
I will look first at signs that the old model is beginning to break down (for example,
the fact that electronic versions of MLDs have begun to include content not present in
the print editions); then consider current challenges and opportunities; and finally
suggest what the MLD might look like ten years from now.

Optimizing the role of language in Technology-Enhanced Learning (2007): 49-51

50 M. RUNDELL

3. Issues and challenges

Probably the most obvious point is that the MLD may no longer be quite ‘M’ or ‘D’.
The old binary choices in reference publishing (monolingual or bilingual, dictionary or
encyclopedia, advanced or intermediate) may no longer be relevant. Customization
and personalization are likely new directions, so the current globally-marketed one-
size-fits-all package will probably be unpicked. From the point of view of content,
lexicographers and linguists have never been better placed. The age of data-sparseness
is behind us, and we have fantastic language resources at our disposal (corpora of
infinite size, and language-analysis software of increasing power and sophistication:
e.g. Kilgarriff and Rundell 2002). Effective exploitation (in dictionaries) of learner
corpora has only just begun (Rundell and Granger 2007; Gilquin, Granger and Paquot
forthcoming), and there is much more to be done on this front. Essentially, we can do
anything, and there are plenty of areas of the language that dictionaries do not yet
describe adequately.
The challenges include:

matching content to users’ needs: one of the issues here is that so much
reference material is available at no cost on the Web (Google, Wikipedia etc.),
so we have to be clear about what to focus on. The challenge is to work out how
to provide information which learners need, and which is either not available
elsewhere, or not available in an easy-to-use form that takes account of
learners’ needs (and limitations);

how reference data will be delivered: the current platform for electronic
dictionaries is the CD-ROM, already an ageing technology with obvious
limitations, so what (in addition to online access) might replace it? A possible
scenario is to see our reference materials as a set of components which
customers can mix and match according to their needs. For example, a learner
from China doing a Masters in agriculture at a British university could have
access to a ‘core’ ELT dictionary with the option of Chinese translations,
academic-writing aids, and subject-specific terminology. The resources thus
become less static, more dynamic;

the hardest question: how to fund all this development? Electronic versions of
MLDs have been around for 15 years or so, but none have yet made any money
(and they cost a lot to develop). New revenue models need to emerge, and these
could include advertising. (The typical users of MLDs are an adman’s dream:
young, intelligent, aspirational etc.). To be discussed…

4. References
DE SCHRYVER G.-M. (2003), “Lexicographers’ dreams in the electronic dictionary age”, in
International Journal of Lexicography, 16.2: 143-199.
GILQUIN G., GRANGER S. and PAQUOT M. (forthcoming), “Learner corpora: the missing link
in EAP pedagogy », in Journal of English for Academic Purposes.
 THE DICTIONARY OF THE FUTURE 51

KILGARRIFF A. (2006), “Collocationality and how to measure it”, in E. Corino, C. Marello,

C. Onesti (eds.), Proceedings of 12th EURALEX International Congress, Alessandria,
Edizioni Dell’Orso: 997-1004.
KILGARRIFF A. and RUNDELL M. (2002), “Lexical Profiling Software and its Lexicographic
Applications – a Case Study”, in A. Braasch and C. Povlsen (eds.), Proceedings of the 10th
EURALEX Copenhagen Proceedings: 807-818.
RUNDELL M. (1998), « Recent trends in English pedagogical lexicography”, in International
Journal of Lexicography, 11.4: 315-342.
RUNDELL M. and GRANGER S. (2007), “From corpora to confidence”, in English Teaching
professional, 50 : 15-18.
SINCLAIR J.M. (ed.) (1987), Looking Up: the COBUILD project in lexcial computing,
London, Collins ELT.
 Macmillan English Campus: A case study

Emma Shercliff
MacMillan English Campus

1. Introduction
Macmillan English Campus is an online practice environment designed for the
learning and teaching of English as a Foreign Language. It was developed in
conjunction with one of the world’s leading language schools, Cultura Inglesa, Sao
Paolo, and is today being used by over 90,000 students worldwide.
Macmillan English Campus consists of two components:
1. A flexible database of over 3,000 highly interactive language activities, developed
by Macmillan's leading ELT authors. These activities include interactive language
exercises, listening tasks, pronunciation exercises, vocabulary exercises, progress
tests, exam preparation exercises, language games, web projects and weekly news
items. All users also have access to an online version of the Macmillan English
Dictionary.
2. Sophisticated content management software, allowing institutions to manage their
users and chart our online resources to their own courses and course materials. The
Macmillan English Campus platform includes an electronic mark book and
personalisation tools for each user.
The concept behind the Macmillan English Campus is that language learning can be
greatly enhanced by an effective combination of face-to-face teaching and customized
online support materials. It is this blended learning solution that makes the Macmillan
English Campus unique. It ensures that our users continue to receive face-to-face
tuition and contact with their teachers whilst remaining free to study online within a
controlled learning environment.

2. Outline of the presentation/demo

Macmillan English Campus, an online language learning environment, is at the cutting
edge of TEL developments. It makes use of exciting new web technologies to deliver
an innovative learning experience for both teachers and students. As such, it provides
an excellent example of the challenges encountered by practitioners on a daily basis
when exploring new and more effective ways of dealing with language.

Optimizing the role of language in Technology-Enhanced Learning (2007): 53-55

54 E. SHERCLIFF

This presentation will comprise a case study of the Macmillan English Campus, which
was launched in 2003. I will outline the concept behind the ‘blended learning’
pedagogy of the Macmillan English Campus and will then address the issues and
challenges we have faced over the past four years, with specific reference to the
experience of language teachers wishing to integrate technology-enhanced learning
into their teaching programmes for the first time. I will outline the enhancements we
have incorporated into the Macmillan English Campus learning platform as a result of
user feedback and outline future developments we have planned for 2008 and beyond.
In the light of our extensive experience developing online learning applications, I will
also highlight what we believe to be the limitations of technology in a language
learning context.
I will give specific examples of TEL methods and tools and demonstrate some of the
new functionality, such as teacher-to-student messaging, recently incorporated into the
Macmillan English Campus.
The Macmillan English Campus has been adopted by a number of teaching prestigious
institutions, schools and universities worldwide, including the International House
World Organisation, the British Council and the Bell Schools network. The
presentation will draw on Macmillan English Campus’s widespread experience in the
field and is intended to focus on practice rather than theory. Much of our publishing is
driven by user responses to our learning platform and we have therefore developed
sophisticated mechanisms for gathering and evaluating feedback from users across five
continents.
By sharing the experiences of Macmillan English Campus, I will offer a practical
insight into the challenges of developing materials for technology enhanced language
learning which I hope will stimulate comment and debate.

3. Issues and challenges

The issues and challenges faced at Macmillan English Campus include the following:
Publishing
• How to publish for the web – inventing a new authoring tool
• Ensuring the English Campus is tailored to a school’s teaching programmes and
pedagogic style to enable true ‘blended learning’
• Digital asset management
• The limitations of technology e.g. speech recognition tools

Commercial
• Investment: enormous cost of developing online platform
• Perception amongst certain customers that digital product should be free
• Initial assumption amongst certain customers that teachers could author their
own material at lower cost
 MACMILLAN ENGLISH CAMPUS: A CASE STUDY 55

• Decision making process slow: initial reluctance of institutions to embrace

online learning platform as adoption necessarily involves a change in pedagogy
and teaching methodology

Training & Support

• Teacher training
• Lack of specific technical expertise within language learning organisations
• Creating an online community for Macmillan English Campus users to share
examples of best practice (daily blog now available at
www.macmillanenglishcampus.com/support)
 The Base lexicale du français (BLF), a free web-
based learning environment for French vocabulary

Serge Verlinde
Katholieke Universiteit Leuven

1. Introduction
This presentation deals with recent developments in web-based electronic learner’s
dictionaries and their use in CALL (computer assisted language learning) applications.
My research interests are the lexicon and its structure, corpus analysis and CALL. I am
coauthor of the Dictionnaire d’apprentissage du français des affaires (DAFA) and
have developed, together with Thierry Selva, the Base lexicale du français
(www.kuleuven.be/ilt/blf), a free accessible learning environment (online dictionary
and exercises) for French vocabulary.

2. Outline of the presentation/demo

Today electronic dictionaries, and electronic pedagogical dictionaries in particular, are
much more than an electronic version of a paper dictionary. Combined with NLP
applications, they may be turned into a powerful language teaching/learning (and
research) tool.
The BLF is a free web-based learning environment of a new generation, which
combines
-- a learner’s dictionary or lexical database (Dictionnaire d’apprentissage du français
langue étrangère ou seconde – DAFLES);
-- a corpus of newspaper texts;
-- a CALL application (Alfalex);
-- direct access to other freely accessible lexical resources on the web.
The BLF was developed from scratch and it is based on a relational database.
Theoretically, and as far as the data-structuring allows, one should be able to launch
any query on a lexical database. These queries could apply to the lexicon itself
(nomenclature, word combinations) as well as to its properties (e.g. grammatical
category for the nomenclature; lexical function for the word combinations), and even

Optimizing the role of language in Technology-Enhanced Learning (2007): 57-59

58 S. VERLINDE

to a series of characters contained in a cell of the database (e.g. a query concerning all
definitions encompassing the noun action or all verbs used with a prepositional group
introduced by the preposition à). In the BLF, we have tried to reach these goals within
the didactic perspective of teaching/learning French as a foreign language, as well as
exploit these resources for research purposes.
The corpus is used to provide both examples of the use of multiword units (word
combinations) and sentences for the exercises in the CALL application (ALFALEX)
by using NLP-tools.
ALFALEX offers about ten different types of exercises relating to 'the words’ most
important features:
- formal features (morphology, verb conjugation, derivation);
- intrinsic features (gender);
- combinatorial features (use of prepositions after verbs, nouns and adjectives,
multiword units;
- lexical relations (synonyms, schémas actanciels or words encountered in the same
communicative situation: e.g. how do we designate the act of killing (assassiner) a
person ? un assassinat ; what do we call the person who killed another person ? un
assassin ; and the person who was killed? la victime);
- translation (decoding: French > Dutch, encoding: Dutch > French).
The exercises listed above are semi-automatically generated through direct use of the
information in the lexical database and the corpus (for the contextual exercises).
Directional and constructive feedback is provided: by means of hyperlinks, the user
can access the lexicographical description, which is available for almost every item in
the exercises. Twice a year, ALFALEX also automatically generates a qualitative
report for every user of the environment.
Unfortunately, non-commercial dictionaries such as the DAFLES only cover a part of
the lexicon. Therefore, if a user submits a word which is not listed, he will be
redirected to other lexical resources available on the internet. He also has access to
other free web resources (e.g. corpora, semantic networks) for French.

3. Issues and challenges

-- How do learners use electronic dictionaries and the most recent resources available
on the web? How can the dictionary be tailored to users’ real needs? Tracking and
logging the real use of (web-based) electronic dictionaries is certainly the first step
of discovering this.
-- Learners still have problems using the dictionary for decoding purposes (length and
structure of the entries, incomplete lexical description, identification of multiword
 THE BASE LEXICALE DU FRANÇAIS (BLF) 59

units, …). Could NLP applications (e. g. the use of a parser), combined with a
dictionary/lexical database, be helpful?
-- Encoding is even more complicated. How can we develop a real writing assistant?

4. References
ABEL A. and WEBER V. (2000), ELDIT – A prototype of an innovative dictionary, in U. Heid
et al. (eds), Proceedings EURALEX, The Ninth EURALEX International Congress,
Stuttgart : 807-818.
ALDABE I., ARRIETA B., DÍAZ DE ILARAZZA A., MARITXALAR M., NIEBLA I., ORONOZ M. and
URIA L. (2006), “The use of NLP tools for Basque in a multiple user CALL environment
and its feedback”, in P. Mertens, C. Fairon, A. Dister and P. Watrin (eds), Verbum ex
machina. Actes de la 13e conférence sur le Traitement automatique des langues naturelles,
Louvain-la-Neuve : 815-824 (Cahiers du Cental 2).
ANTONIADIS G., ECHINARD S., KRAIF O., LEBARBÉ T. and PONTON C. (2005), « Modélisation
de l'intégration de ressources TAL pour l'apprentissage des langues : la plateforme
MIRTO”, in Alsic.org, vol. 8: 65-79.
ANTONIADIS G., ECHINARD S., KRAIF O., LEBARBÉ T., LOISEAU M. and PONTON C. (2004),
“CALL: from current problems to NLP solutions MIRTO: a user-oriented NLP based
teaching platform”, in Proceedings of EuroCALL Conference 2004, Vienna.
BLUMENTHAL P. (2006): Wortprofil im Französischen, Tübingen.
BLUMENTHAL P. and HAUSMANN F.J. (eds.) (2006), Collocations, corpus, dictionnaires, in
Langue française, 150.
DE SCHRYVER G.-M. (2003), Lexicographers’ Dreams in the Electronic-dictionary Age, in
International Journal of Lexicography, 16.2: 143-199.
GAUME B. (2004), “Ballades aléatoires dans les Petits Mondes Lexicaux”, in: I3, Information
Interaction Intelligence, 4.2 : 1-59 (w3.univ-tlse2.fr/erss/textes/pagespersos/gaume/
resources/I3.impression.5.pdf).
GROSSMANN F. and TUTIN A. (eds.) (2003), “Les collocations. Analyse et traitement”, in
Travaux et recherches en linguistique appliquée, série E, n° 1.
HERBST T. and POPP K. (1999), The Perfect Learners’ Dictionary (?),Tübingen.
RUNDELL M. (1998), “Recent trends in pedagogical lexicography”, in: International Journal
of Lexicography, 11.4: 315-342.
VERLINDE S., SELVA T. and BINON J. (2005), “Dictionnaires électroniques et environnement
d'apprentissage du lexique”, in Revue française de linguistique appliquée, X.2: 19-30.
VERLINDE S., SELVA T. and BINON J. (2006), “The Base lexicale du français: a
Multifunctional Online Database for Learners of French”, in Corino E., Marello C., Onesti
C. (eds.), Proceedings XII Euralex International Congress. Torino, Italia, September 6th-
9th 2006, Torino, vol. II: 471-481.
VERLINDE, S., BINON J., OSTYN S. and Bertels A. (to appear), “La Base lexicale du français
(BLF): un portail pour l’apprentissage du lexique français”, in Cahiers de Lexicologie.
 Linguistic anomaly

Carl Vogel
Trinity College Dublin

1. Introduction
My work in computational linguistics is influenced by the course of my education: a
liberal arts undergraduate degree in computing, literature, philosophy and psychology;
an MSc by research in artificial intelligence focussed on inheritance reasoning for
constraint based syntax; a PhD in cognitive science on models of default reasoning and
their relation to human reasoning.

2. Outline of the presentation/demo

The direction of research is attuned to cognitive science paradigms, within the field of
computational linguistics. Thus, I am interested in formal language theory from the
perspective of, for example, theoretical learnability results more than the impact on
engineering artifacts, although I am also interested in those. The Irish language
spelling checker licensed by the most ubiquitous software company is an example.
From my perspective, the practical value was in providing support for daily use of the
Irish language in working settings taken for granted for immediate linguistic support
and feedback by speakers of the world’s major languages, and the theoretical interest
was in devising both a space-efficient data structure (think of it as an optimized two
pointed trie – something like a rugby ball more than a trie) for a large wordlist, and a
time-efficient method of populating it.
This has turned into a focus on linguistic anomaly. However, the cognitive science
background entails empirical and theoretical interest, in general.
While I have done work in descriptive syntax within Head-Driven Phrase Structure
Grammar (HPSG), particularly on quirky case, I have also worked on parsing for
HPSG and formal foundations of HPSG attending to the feature logic. In particular, I
am interested in paraconsistent feature logic for a version of HPSG that affords
description of degrees of grammaticality. In this context, the related work with and by
Jennifer Foster has been useful.
With respect to semantics, like many, I have addressed underspecification – compact
representations of ambiguity that model the modest human overload that comes with

Optimizing the role of language in Technology-Enhanced Learning (2007): 61-63

62 C. VOGEL

ambiguous language, yet the impressive facility humans have for reasoning with only
partial resolution of ambiguity. However, I have also considered “overspecification”
the process by which accepted linguistic expressions have their senses extended to
new meanings and the constraints that exist both theoretically, and in human behavior,
with respect to sense extension. This is intimately linked to metaphoricity.
I have also studied study of logics of human reasoning with defaults, chains of
statements that express regularities confronted with exceptions. Here there are
concerns with the formal properties of the logics themselves, and with the degree to
which they serve as adequate models of human reason with generalizations that have
exceptions.
Thus, my interests in linguistic anomaly span from orthographical well-formedness,
through appropriateness of lexical meaning, formal syntactic description and degrees
of grammaticality, to semantic well-formedness and sense extension for
representations, and reasoning with incomplete and inconsistent information.
Currently, my funded research is in techniques for text classification, looking
particularly at linguistic change over time, towards establishing milestones of normal
language development and decline, particularly addressing the Iris Murdoch corpus as
a source of data that may reveal features that correlate with progression of Alzheimer’s
disease.
The uniting theme in all of these sorts of linguistic anomaly that I study is the tension
between linguistic convention and linguistic creativity: ill-formedness versus
creativity.

3. Issues and challenges

The practical challenge for work on semantics and reasoning is in advancing beyond
spelling checkers and text entailment contests to semantic checkers, tools that can be
used to improve drafting of legislation and support human decision making.
The main empirical challences are those shared with psychologists and
psycholinguistics generally, in defining experiments that meaningfully test theoretical
claims.
Large empirical challenges are also associated with corpus linguistics: the longitudinal
analysis of language change and corelations between milestones of language change
and life events depends on a kind of corpus collection that simply does not adequately
exist at the present.

4. References – Some Representative Publications

APPEL C. and VOGEL C. (2001), “Investigating Syntax Priming in an E-Mail Tandem
Language Learning Environment”, in K. Cameron (ed.), C.A.L.L. – The Challenge of
Change: Research and Practice, Elm Bank Publications, Exeter: 177-184.
 LINGUISTIC ANOMALY 63

DEVITT A. and VOGEL C. (2003), “Using Wordnet Hierarchies to Pinpoint Differences in

Related Texts”, in Proceedings of EUROLAN 2003: Ontologies, and Information
Extraction International Workshop: 37-44.
EBERLE K. and VOGEL C. (2000), “Compact Representations of Ambiguous Language”, in
N. Nicolov and R. Mitkov(eds), Recent Advances in Natural Language Processing II, John
Benjamins: 191-206.
FOSTER J. and VOGEL C. (2004). “Parsing Ill-Formed Text Usin gan Error Grammar”, in
Artificial Intelligence Review, 21: 269-291.
HEALEY P.G.T., VOGEL C., and Eshghi A. (2007), “Group Dialects inan Online Community”,
in R. Arnstein and L. Vieu (eds), DECALOG 2007,The 10th Workshop on the Semantics
and Pragmatics of Dialogue, Universita di Trento, May 30 - June 1, 2007: 141-147.
POPOWICH F. and VOGEL C. (1991b), “A Logic Based Implementation of Head-Driven Phrase
Structure Grammar”, in C. Brown and G. Koch (eds), Natural Language Understanding
and Logic Programming, III, Elsevier, North-Holland: 227-246.
SCHOETER A. and VOGEL C. (eds), (1995), Nonclassical Feature Systems, Vol. 10 of
Edinburgh Working Papers in Cognitive Science, University of Edinburgh, 216 p.
VAN GIJSEL S. and VOGEL C. (2003), “Inducing a Cline from Corpora of Political
Manifestos”, in M. Aleksy et al. (eds), International Symposium on Information and
Communication Technologies : 304-310.
VOGEL C. (2001), Dynamic Semantics for Metaphor, in Metaphor and Symbol, 16 (1 & 2):
59-74.
VOGEL C. (2007), “N-gram Distributions in Texts as Proxy for Textual Fingerprints”, in A.
Esposito, E. Keller, M. Marinaro and M. Bratanic (eds), The Fundamentals of Verbal and
Non-Verbal Communication and the Biometrical Issue, IOS Press: 189-194.
VOGEL C. and POPOWICH F. (1997), “A Parametric Definition for a Family of Inheritance
Reasoners”, in New Generation Computing, 15: 247-292.
VOGEL C. and VILLADA B. (2000), “Spanish Psychological Predicates”, in R. Cann, C. Grover
and P. Miller (eds), Grammatical Interfaces in Head-driven Phrase Structure Grammar,
CSLI Publications, Stanford: 251-66.
VOGEL C. (1996), “Human Reasoning with Negative Defaults”, in D. Gabbay and
H.J. Ohlbach (eds), Practical Reasoning: 606-621. Lecture Notes in Artificial Intelligence
1085, Springer Verlag, Berlin. Proceedings of the International Conference on Formal and
Applied Practical Reasoning, FAPR’96, Bonn, June 1996.
 Evolving approaches to web-supported
language learning:
From platforms to platform-independent tools

David Wible
National Central University

1. Introduction
The task of designing systems and tools that support language learning on the Web is
changing due to evolving demands on such technologies from various sources. The
design of Learning Activity Management Systems (LAMS) dedicated to language
learning, for example, is facing a countervailing trend toward the use of all-purpose
platforms such as Blackboard and Moodle. Parallel trends toward the consolidation of
systems and content creation include standards specification movements such as
SCORM. This overall convergence of content standards and a few all-purpose
platforms does not represent an unqualified positive benefit for language pedagogy.
The advantages are limited by the unique nature of language learning among learning
domains and by the growing availability of digital language tools that ignore both
SCORM conformity and portability into larger platforms. Finally, the communicative
turn in language pedagogy highlights the need for individualized learning experiences
suited to each learner’s communicative needs and interests. This sort of
individualization is traditionally the forte of ITS (Intelligent Tutoring Systems), yet
few of these focus on language learning; those that do are stand-alone systems having
virtually no interoperability with other platforms; they treat narrowly-defined
dimensions of language learning; and they do not extend or scale up easily or at all.
In this talk, I describe some ongoing work by our team in Taiwan that addresses these
current challenges in the design of Web-supported language learning technologies.

2. Outline of the presentation/demo

The first part of the presentation sketches a dedicated language learning platform that
has been created and implemented in Taiwan over the past 8 years. IWiLL is the most
widely used language learning platform in Taiwan (http://www.iwillnow.org). Since
2000, it has been used by more than 190 schools in Taiwan, over 600 teachers, more
than 20,000 students. As an online environment, it provides authoring tools for

Optimizing the role of language in Technology-Enhanced Learning (2007): 65-67

66 D. WIBLE

interactive, multimedia web-based language activities and lessons. It tracks and

profiles learner activity and teacher feedback, and automatically stores learners’
writing in a dynamic learner corpus (over 3 million words indexed with over 70,000
tokens of teacher feedback).
The second part of the presentation describes a recent collaboration between the
IWiLL design team in Taiwan and City University of Hong Kong aimed at
modularizing the writing components of IWiLL, creating building blocks or modules
from IWiLL’s unique language teaching and learning functions so they can be
imported into City University’s Blackboard platform.
Part three describes a further step in this diffusion away from a dedicated language
learning platform toward flexible, modular tools. This step involves browser-based
language learning tools which are completely platform independent and accompany
learners on their unrestricted navigation of the Web. These ubiquitous tools detect
linguistic features in real time on the web pages that the user freely browses and
discretely offers these for the selective attention of the learner. This approach is
illustrated with a tool designed by our team in Taiwan that focuses on collocations,
called Collocator.

3. Issues and challenges

The main issue in this research is how to create technologies that have high portability
and integrate well in existing Web environments. The most portable of these novel
technologies face the additional challenge of being required to perform in real time
under noisy, unscripted conditions on the Web. These challenges result from our
approach, which contrasts with more widely known stand alone language tools that do
not integrate well within existing Web environments or that rely on prescribed content
or fragile and baroque learner models that do not scale well.

4. References
IKEDA M., ASHLEY D.K. and CHAN T.-W. (eds), Intelligent Tutoring Systems: Eighth
International Conference, ITS 2006, Lecture Notes in Computer Science 4053, Springer,
Berlin.
WIBLE D. (in press), “Multiword Expressions and the Digital Turn”, in F. Meunier and
S. Granger (eds), Phraseology in Language Learning and Teaching, John Benjamins,
Amsterdam.
WIBLE, D. (2005), Language Learning and Language Technology: Toward Foundations for
Interdisciplinary Collaboration, Crane, Taipei.
WIBLE D., KUO C.-H., CHEN M.C., TSAO N.-L. and HUNG T.-F. (2006), “A Ubiquitous Agent
for Unrestricted Vocabulary Learning in Noisy Digital Environments”, in Lecture Notes on
Computer Science, 4053: 503-512.
WIBLE, D., KUO C.-H., TSAO N.-L. (2004), “Contextualizing Language Learning in the
Digital Wild: Tools and a Framework”, in Proceedings of IEEE International Conference
on Advanced Learning Technologies, Joensuu.
 EVOLVING APPROACHES TO WEB-SUPPORTED LANGUAGE LEARNING 67

WIBLE D., KUO C.-H., TSAO N.-L., HSIU-LING LIN A.L. (2003), “Bootstrapping in a
Language Learning Environment”, in Journal of Computer-Assisted Learning, vol 19 #1:
90-102.
 The SACODEYL project – Corpus exploitation
for language learning purposes

Johannes Widmann
University of Tübingen

1. Introduction
SACODEYL is situated in the field of computer-assisted language learning with the
help of recent developments in corpus research.
SACODEYL is a project within the SOCRATES-MINERVA initiative whose main
aim is to develop an ICT-based system for the assisted compilation and open
distribution of European teen talk. This scheme encompasses two groups: group 1,
youngsters between 13 and 15 and, group 2, those between 16 and 18.
The main aim of the project is for young Europeans to use corpora for the learning of
languages. The pedagogical rationale of the project rests upon notions of autonomous
learning and meaningful interaction. SACODEYL users will come into close contact
with the real voices of peer young Europeans from other countries, their feelings,
opinions and speech, without the mediation, otherwise natural and necessary, of third
parties such as publishing houses. These peer group voices will make it easier for
young people to identify with the language and the contents being taught.
The SACODEYL project aims inter alia at developing a pedagogically-driven search
tool for querying corpora in such a way that allows language teachers to access
corpora from their teaching perspectives and experiences. The basis of the corpus
annotation will be the SACODEYL annotation and corpus enrichment scheme that
will be used with the raw transcripts.

2. Outline of the presentation/demo

First, I would like to show the design of our annotation tool and the rationale that has
been guiding its development. The novel and most important aspect of this annotation
tool as opposed to existing tools is the fact that it is based solely on pedagogical
criteria.
On the one hand, this includes traditional grammatical and lexical annotation
categories that are based on pedagogical grammars. On the other hand, this includes
newer categories, such as communicative functions, references to various CEF scales,

Optimizing the role of language in Technology-Enhanced Learning (2007): 69-71

70 J. WIDMANN

information on typical spoken language properties, and information on texture

properties. Teachers will be able to modify the corpus annotation and thus they can
include their own information in any way they wish. The important thing is that they
won’t need any advanced computational knowledge. The annotator produces standard
TEI-conform XML files without the user having to know about XML. During the
demo session, we will look at a beta version of the annotation tool and also at some
annotated interviews that have been produced in the project.
Second, we will be able to look at a prototype of the search tool which can be used to
retrieve the data and the annotation of the corpora. This online tool will be available to
all interested parties free of charge. The tool provides for different viewing
perspectives of the corpus. In contrast to many existing query tools, it provides not
only a concordance-based viewing perspective but also a section-based viewing
perspective where different search options can be viewed in contrast, such as co-
occurrences of words and annotated categories within the same section.
To our knowledge, this is one of the first projects where corpus linguistic
methodologies have systematically been matched with pedagogic criteria and where
pedagogic criteria have been essential in the design of the software. The project
explicitly takes into account teachers’ needs and perspectives and it aims at offering a
low threshold for those teachers who are not very much acquainted with ICT. All the
tools are menu-driven. The search tool is an online jsp program requiring no previous
installations.

3. Issues and challenges

Given the feedback that we have received so far, it seems as if the SACODEYL
project is filling a gap with its approach that has not yet been adequately addressed.
We have had quite a number of teaching sites responding positively to our design.
Some of the challenges that lie ahead can be phrased as follows: So far, very little
computational intelligence has been included in the software of the project. All the
annotation has to be done by hand. On the one hand, this is positive as it is less
daunting for the teachers to start off with corpus-based teaching. On the other hand,
the project does not yet take advantage of some of the possibilities that computational
linguistics seems to offer. This remains a challenge that asks for closer cooperation
between pedagogically-interested linguists and technology-interested linguists.
Furthermore, the design of the web-based management and integration of the corpora
is also an issue that needs more attention. We have designed an integrated system, but
due to a lack of resources it has not yet been possible to fully implement the system.

4. References
BRAUN S. (2005), “From pedagogically relevant corpora to authentic language learning
contents”, in ReCALL 17:1: 47-64.
 THE SACODEYL PROJECT 71

BRAUN S. (2006), “ELISA – a pedagogically enriched corpus for language learning

purposes”, in S. Braun, K. Kohn and J. Mukherjee (eds), Corpus Technology and
Language Pedagogy: New Resources, New Tools, New Methods, Peter Lang, Frankfurt/M.
MUKHERJEE J. (2004), “Bridging the gap between applied corpus linguistics and the reality of
English language teaching in Germany”, in U. Connor and Th. Upton (eds), Applied
Corpus Linguistics: A Multidimensional Perspective, Rodopi, Amsterdam.