Abstract:

In Bangladesh, majority students outside Dhaka specially the rural students struggle to pass in English in the national examinations held in different levels, Primary School Certificate (PSC) (5th grade), Junior School Certificate (JSC) (8th grade), Secondary School Certificate (SSC) (10th grade), Higher Secondary School Certificate (HSC) (12th grade), and equivalent examinations from Madrasa and Technical Education Board. Overall result is greatly undermined by poor scores in English examination. As the result shows, there is a great lack of English skills. This gap greatly increases the difficulty for the students for their next stage in academic and professional life. It is also withholding the whole nation’s progress. If English education can be slightly improved, it will have huge impact on the nation. Rural students are falling far behind in countrywide tests due to lack of quality and extra-care enjoyed by urban students.

With millions of candidates appearing in each of the examinations, the sheer volume of the students involved makes the problem challenging. As a chicken and egg problem, there is a scarcity of English teachers to cater the target audience. Moreover, many of them are not well-trained for effective English teaching. Training teachers from one village at a time will take too long to achieve. Additionally, the extra-care enjoyed by urban students like private tuition, surrounding atmosphere which uses English a lot would be too expensive. Replicating urban environment in rural areas would be infeasible. Audio-visual aids in classroom is often unavailable. Due to lack of exposure, both teachers and the students have significantly small vocabulary. As a result, students are forced to use dictionary to comprehend the textbook and to prepare for examinations. Reduced vocabulary makes use of mono-lingual dictionary to learn new words impractical. Bi-lingual dictionary is often generic in nature, does not give meaning in all senses, does not indicate which of the given meanings is most appropriate for the sentence the learner is trying to understand or construct, vocabulary set may not sufficiently overlap, etc. As language is always evolving, meanings of words vary from time to time. A learner needs to know how a word or phrase is used in current times, in given context. Regular dictionary fails to serve this purpose.

No empirical research has been conducted with the English text book of national examinations to help the students in the rural areas of Bangladesh. We choose to build COBUILD dictionary to expose learners to usage in complete sentences instead of extracted phrases. Viewing multiple example sentences demonstrating almost every sense of every word in actual usage would assist in both comprehension and construction of sentences. As a proof of concept, we choose to start with SSC English first paper textbook. This is the first attempt at computational lexicography to build the learners dictionary for the target audience. In this paper, we present the initial heuristics and algorithms used to extract the information, and present the result/findings of the initial prototype.

Keywords:

Technology-enhanced learning, corpus, computational linguistics, cobuild dictionary, learner dictionary, SSC English, rural English.