Nation (1982, 1990, 1994) and Meara (1980, 1988, 1993) have argued for
years that learners are lexically under-challenged in the majority of published
courses. Barnard (1971) argued that students could learn 2000 words in five
years of study; Meara (1980) argued that they were probably capable of learning
more like 2000 a year. The present study suggests that Meara was closer,
and explores one way of building up volume without sacrificing depth. The
learner-as-lexicographer fiction and technology allows large numbers of
words to be met and processed in context.
Recapitulation
In the introduction, several desiderata were set out for a second-language
reading tutor: The tutor should be extensively used by a large number of
students over a lengthy period; it should be integrated into an ongoing
curriculum; it should be based on theories deriving from basic research;
it should be tested for learning effectiveness against a control group,
and this information fed back to the development process; it should involve
the reading of extended texts; it should use the computer to do things with
text that cannot be done or easily done on paper; it should invite students
to ask rather than answer questions. It was proposed that the concordance
concept and technology made it possible to group these desiderata within
an extended program of software development.
Large numbers and lengthy period: As of June 1996, nearly 1000
students at SQU have used some form of corpus tutor in some way over three
years, some for as many as 30 hours. PET·2000 is probably one of
the ten hardest-used tutors in the history of CALL, to the extent that
amount of use gets reported in the literature.
Curriculum integration: The concordance tutor was tightly integrated
into an ongoing curriculum. It was based on clearly analysed student needs
which were impeding progress in their studies, in some cases forcing them
to abandon their studies. All computer work was reinforced within the same
week in the classroom.
Based on theory and research: The tutor was designed to exploit
research-based information and test its instructional applicability. Instructional
research has shown that transferable word knowledge is mainly produced
by extensive reading. Both instructional and psychological research have
isolated the mechanism of transfer in multicontextuality. This study tested
and confirmed the hypothesis that some of the benefits of multicontextuality
can also be realized with a concordance program.
Feedback to development: Learning effectiveness was tested at
every step and fed back to development. Particularly, the lessons of PET·200
were fed into the development of PET·2000. The basic multi-contextuality
principle was validated in PET·200, applied to massive acquisition
in PET·2000, and will be further tested against the University Word
List pending the development of a suitable corpus.
Something to read: PET·2000 involved the reading of texts,
possibly as many as 200 paragraphs a week, about 20 pages, for heavy users.
Exploiting the computer: The computer was used to present text
in ways that cannot be done practically on paper. Assembling all examples
of a word throughout a term's reading is routine for a concordance, next
to impossible on paper or in a classroom. Also, the computer helped the
learners produce high-quality, multi-page documents that they could hardly
create by themselves (the glossaries), but which were nonetheless created
as a result of their inputs and decisions.
Asking not answering: Many students used PET·2000 to
ask for 1000 or more examples of words in use, probably more questions
than they asked in class plus all their dictionary look-ups combined.
Whatever happened to the PET?
The reader may wonder why the PET itself has not been used as one of
the dependent measures in this study. The reason involves University policy.
After PET testing, students and instructors alike are informed of band levels
but not raw scores. Since band level is hardly a fine measure of learning,
as is particularly needful in the case of vocabulary and reading, it was
not used in this study. The PET was used rather to derive an experimental
learning task and motivate subjects.
Still, the claims for corpus tutoring would be weakened if the learning
detected in this study had not affected the Omani students' PET experience
in some way. In fact, the PET has effectively disappeared as a major problem
in the College of Commerce, although there is no way of disentangling all
the contributing factors. In May, 1994, only 5 out of 20 Band 3 students
cleared the Band 4 hurdle; in December, 1995, all 17 Band 3 students went
to Band 4; and similar trends existed at the lower levels.