Computational Linguistics and its Use in Real World: the Case of Computer Assisted-Language Learning

Michael ZOCK
Langage & Cognition
LIMSI-CNRS, B. P. 133
91403 ORSAY - FRANCE
e-mail:zock@limsi.fr

Surprising as it may be, one of the biggest markets
for products of computional linguistics (CL) has been
largely overlooked: the classroom. While machine
translation has attracted a considerable amount of
research, hence resources, CALL, ~ a domain with a
comparable potential, has hardly ever received the
attention it deserves. Actually there seems to be a
communication problem and a mutual lack of interest
concerning the work done in the neighbouring
disciplines.
Computational linguists don't show much interest
for CALL, and CALL experts ignore the work done
by computational linguists. Strangely enough, even
within the ITS, CAI, CBI OR ICAI communities,
little, if any reference is made to work done in CALL
(4, 10, 13, 15, 18, 24, 25, 26, 28, 30, 33, 35). This
being so, it is hardly surprising to see that the domain
is never mentioned in textbook on Artificial Intelligence
or Psycholinguistics. Yet there are a number of
publications in psychology that deal with related
issues such as learning theory (12), language
learning (5, 6, 11, 14, 20), language teaching (6, 16,
19, 27), educational technology, i.e. programmed
instruction (9, 12, 31), theory of writing (3),
algorithmization of the learning process (17),
learning strategies, i.e. learning how to learn (23),
etc.
It is also worth mentionning that no cross fertilization
has taken place between the CALL community
and people working in the Machine Learning
paradigm (7, 21). 2 While there are fundamental
differences in terms of goals and methods, there are
also some important overlaps. Books on more
sophisticated CALL systems are still scarce (13, 32),
so is the work that shows how current NLP
technology could be used in the classroom (1, 22, 34)
Yet CALL is a field with considerable potential. It
is both a challenge and a chance to bring NLP
' Unlike 17S (Intelligent Teaching Systems), CAI (Computer
Assisted Instruction), CBI (Computer Based Instruction)
and ICAI (Intelligent Computer Aided Instruction),
which use language ['or communicating domain specific
knowledge, CALL has langage learning as its primary goal.
Obviously, NLP-technology may be relevant for all these
systems, but in different ways.
For a slightly outdated bibliography on CALL, see (2).
technology from the research laboratories to the real
world. While computational linguists will certainly
have to play an important role in providing linguistic
resources (grammars, lexicon) and processing tools,
it is not clear yet how to decide on the adequacy of
the tools (browser, editors). Also, there are good
chances that within this context new problems arise,
while old solution turn out not to be good at all, in
which case the following two questions arise: what is
the nature of these new problems?, and in what terms
do these new problems have to be rethought? Other
related issues of interest are the following:
• what can current NLP technology contribute to
computer-assisted language learning?
how can this technology meet the demands of
pedagogical theory for communicative language
teaching in a natural environment?
what can NLP-based systems teach us about
language acquisition, linguistic theory and
NATURAL language processing in general?
what effect can a domain like CALL, or the
involved disciplines have on the development of
NLP technology?
• what lessons have been, or can be learned by
looking at the available CALL systems?
In order to get a clearer picture of these problems,
and in order to draw the community's attention to the
fact that there is a REAL need and potential for
integrating NLP technologies in CALL systems, we
propose a panel discussion between specialists in the
concerned disciplines (linguistics, artificial
intelligence, psychology, language teaching). The
expected results of such a discussion are not only an
increase of resources (manpower) in the CALL
domain, but also an increase of awareness, that is, a
sharpening of the researchers' understanding of what
the problems are that people encounter when
processing language. All too often we look at
language only from the point of view of the machine,
i.e. how can languages be processed by computers. In
doing so we tend to forget the obvious : natural
languages are used by people.
1002
In building CALl, systems we will realize that
there are ninny problems in the area of natural
language processing that have been either
overlooked, or been posed in inadequate terms. Yet,
if we really want to get a real tmderstanding of the
fimctioning of natural languages, --how they are
used, how they are learned?-- we have to look at the
constraints of the system for which they have been
designed: man. This is the price we have to pay if we
want to produce programs that are of interest not only
in the research labs but also into the arena of real
world.
Strangely enough, in the past we had neither the
right tools, nor a decent theory (see 8, 27, 31), yet
people were optimistic and went ahead. Today we are
much better off. We do have very powerfid tools (fast
computers with well designed graphical interfaces,
browsers, CD-Roms, authoring languages), and a
whole set of quite promising theories, yet we hesitate.
But, what are we waiting lot'?
References:
Abeill6, A. 1990. A Lexicalized Tree Adjoining
Grammar for French and its Relevance to Language
Teaching. In M. Swartz & M. Yazdani (eds.)
Bailin, A, C. Chapelle, L. Levin, G. Mulford, (2
Neuwirth, A. Sanders, R. Sander & ,I. Underwood.
1989. A Bibliography of Intelligent Computer-Assisted
Lauguage Instruction. Computers and the lhunanities.
85-90. Kluwer Academic Publishers
Bereiter, C. & M. Scardamalia. 1987 The Psychology
of Written Composition. Hillsdalc, NJ: Lawrence
Erlbaum
4. Bierman, D., J. Breuker & J. Sandberg (Eds.) 1989.
Artificial Intelligence and Education. Proceedings of
the 4th Intenrational Conference on AI and Education,
Amsterdam
5. Brooks, N. 1964 l,anguage and l,auguage I,earning.
Chicago, Harcourt Brace & Workl
6. Brown, H. 1987 Principles of Language l,carning and
Language Teaching. Englewood Cliffs, N J: Prentice-
Hall
7. Carbonell, J. (ed.) 1990. Machine Learning: Para-digms
and Methods. Cambrklge Mass., the MIT Press
8. Chastain, K. 1969. The audio-lingual habit learning
theory vs. the code-cognitif learning theory, lleidelberg,
IRAL, 7, 2, 97-107
9. De Cecco, J. (ed.). 1964. Educational Technology:
Readings on Programmed Instruction. New York, l lolt,
Rinehart & Winston
10. Frasson, C. & G. Gauthier (Eds.). 1990 Intelligent
Tutoring Systems. Norwood, N J: Ablex
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
Gethin, A. & E. Gunnemark. 1995. The Art and
Science of l,earning Languages. Intellect books,
Headington, Oxford
Hilgard, E. & G. Bower. 1975. Theories of learning.
Englewood Cliffs, N.J.: Prentice Hall
Ilolland, M., J. Kaplan & M. Sams (Eds.).1995.
Intelligent Language Tutors. llillsdale, NJ.: Lawrence
Erlbaum Associates.
Jakobovitz, L. 1970. Foreign Language Learning: a
Psycholinguistic Analysis of the Issues. Rowley,
Newbury House
Kearsley, G. 1987. Computer Aided Instruction. In,
Shapiro (ed.) Encyclopedia of Artificial lntellige,ce.
Vol. 1. New York John Wiley & Sons
Lade, R. 1969. l,anguage Teaching a Scientific
Approach. New York, ltarper & Row
Landa, L. 1974. Algorithmization in Learning and
Instruction. New Jersey, Euglewood Clifl%
Lawler, R. & M. Yazdani (Eds.). 1987. Artificial
Intelligence and Education. Learning Enviromnents and
Tutoring Systems. Vol. l, 413~427. Norwood, NJ:
Ablex
Lcont'cv, A. A. 1974. Psycholinguistik und Sprachunterricht
(Psycholinguistics and l,anguage Tea-ching).
Stuttgart, Kohlhammer
Leont'ev, A.N. 1981. Psychology and the Language
l.,earning Process. Pergamon Press
Michalski, R., J.Carbonell & T. Mitchell (eds.). 1983.
Machine Learning: an Artificial Intelligence Approach.
Pale Alto, Tioga Publishing
Miller, G. & C. Fellbaum. 199(/WordNet and the Organization
of Lexical Memory. In M. Swartz & M.
Yazdani (cds.)
Novak, J. & 1). Gowin. 1984. Learning how to I,earn.
Ca,nbridge. Cambridge University Press
O'Shea T. & J. Self. 1983. Learning and Teaching with
Computers: Artificial Intelligence in Education.
Englewood Cliffs. NJ: Prentice-Itall
Poison, M. & J. Richardson (eds.). 1988. Founda-tions
of Intelligent Tutoring Systems. ltillsdale, NJ.:
Lawrence Erlbamn Associates.
Psotka, J., Massey, D., Mutter, S. (eds.) 1988.
Intelligent Tutoring Systems: Lessons I,ea,ned.
ltillsdale, NJ.: tmwrencc Erlbamn Associates.
Rivers, W. 1964. The Psychologist and the Foreign
Langnage Teacher. Chicago, University of Chicago
Press
1003
28. Self, J. 1988, Artificial Intelligence and Human Learning.
London, Chapman and Hall Computing.
29. Skinner, B. 1968. The Technology of Teaching. New
York, Appleton Century Crofts
30. Sleeman, D. & J. Brown (eds.). 1982. Intelligent Tutoring
Systems. London, Academic Press
3l. Spolsky, B. 1966. A Psycholinguistic Critique of
Programmed Foreign Language Instruction. Heidelberg,
IRAL, 4, 2
32. Swartz, M. & M. Yazdani (Eds.) 1990. Intelligent
Tutoring Systems for Foreign Language Learning: the
Bridge to International Communication. Berlin,
Springer Verlag
33. Wenger, E. 1987. Artificial Intelligence and Tutoring
systems, Computational and Cognitive Approaches to
the Communication of Knowledge. Los Altos, CA:
Morgan Kaufmann Publishers
34. Wilks, Y. & D. Farwell. 1990 Building an Intelligent
Second Language Tutoring System from Whatever Bits
you Happen to Have Lying Around. In M. Swartz & M.
Yazdani (eds.)
35. Woolf, B. 1988. Intelligent Tutoring Systems: a
Survey. In H. Shrobe & American Association for
Artificial Intelligence (Eds.). Exploring Artificial
Intelligence: Survey Talks for the National Conferences
on Artificial Intelligence, Los Altos, CA:
Morgan Kaufmann Publishers. 1-43
1004